In recent years, the regional economic differences in China have aroused widespread concern. With the expansion of regional economic disparities, China's financial development also shows obvious differences at the regional level (Zhao Wei, Ma Ruiyong, 2006). It is of great significance to study the factors affecting regional financial development for explaining and controlling the regional financial development gap. This paper attempts to make a preliminary attempt to solve this problem. In the traditional theory of economic geography, the reasons for regional differentiation are mainly determined by geographical location (such as the distance from the port) and natural conditions. Traditional economic geography can't explain that some regions with poor geographical position can develop well economically and financially. It is this reason that promotes the rise of new economic geography. The key to the new economic geography is Krugman's increasing returns to scale (199 1). The core idea is that the natural conditions and geographical location are very close to each other, or it may be caused by some accidental factors (such as historical events), resulting in regional differences. At the same time, the new economic geography began to consider the economic, historical, cultural and institutional factors that were not involved in traditional geography (Yeung, 2003). However, policy factors, like traditional geographical factors, are often analyzed as accidental events, and they are regarded as indirect rather than direct impacts on economy and finance (today, Zhao Chen, Minglu, 2006). In China, economic policies and traditional geographical factors often do not indirectly affect economic and financial development, but directly play a role. China's policy of "getting rich first" has played an undeniable and direct role in the development of the eastern region. Therefore, the analysis framework of this paper includes the influence of economic policy factors. Financial geography is developed on the basis of new economic geography, which inherits the characteristics of new economic geography beyond the traditional geographical framework to analyze problems, develops a more comprehensive perspective to analyze problems, emphasizes interdisciplinary research, and studies the development of regional finance from the perspectives of politics, economy, culture and history. From this perspective, traditional geographical factors, new economic geographical factors and policy factors are all included. Jin Hotan-Lin (2004) made a preliminary attempt at empirical research in this field, and constructed an evaluation system of regional financial competitiveness from the comprehensive perspective of financial geography, with its index system covering economy, culture, science and technology, location and other aspects. This paper attempts to construct a framework of financial geography and analyze regional financial development from the perspective of financial geography. Under this framework, the influencing factors of regional financial development are mainly divided into three categories: economic geography, new economic geography and economic policy. Second, the descriptive analysis of regional financial development in China We use the financial correlation rate (FIR) and financial marketization rate (FMR) to measure the degree of regional financial development. Financial Correlation Ratio (FIR) proposed by Goldsmith has been widely used as an index to measure the regional financial development. For a long time, China's state-owned finance has a strong administrative color. In order to better reflect the factors that affect China's regional financial development, we deliberately use another indicator-financial marketization rate (FMR) to reflect the development of non-state-owned finance (Zhou Li, 200 1). FIR is defined as the ratio of the value of all financial assets to the value of all physical assets (that is, national wealth), which is the broadest index to measure the relative scale of financial superstructure. FMR is the ratio of financial assets of non-state-owned financial institutions to national wealth. If S stands for deposit, L stands for loan and FIR stands for financial correlation rate, then the calculation formula is: FIR = (S L)/GDP. This method will also be used to calculate the following financial correlation ratio. The corresponding financial marketization ratio FMR is the ratio of the sum of deposits and loans of non-state-owned banks to GDP (Zhou Li, 200 1). It can also be clearly seen from the table 1 that the regional financial development level in China has changed obviously in time and space, and the change of FMR representing the non-state-owned financial development level is more obvious than that of FIR representing the overall financial development level. We will analyze the influencing factors of regional financial development level change from the perspective of financial geography. Thirdly, in empirical analysis, we use mixed regression model and unobserved effect model to analyze the panel data. In order to better explain the problem, we will model and analyze the overall development level of regional finance and the development level of regional non-state-owned finance respectively. (1) Variables and data According to the previous analysis, the factors affecting regional financial development are divided into three categories: 1. Traditional geographical factors. According to the geographical location and the tradition of regional research, we divide 30 provincial administrative units into three regions: east, middle and west. Set two dummy variables, region2 and region3, to represent the central and western regions respectively. 2. New economic and geographical factors. According to the previous studies of Henderson( 1974), Krugman( 199 1), Jin, Tian Lin (2004) and the characteristics of financial geography, we think that the factors that the new economic geography affects the regional financial development are as follows: (1) industrial spillover effect. It is expressed by the proportion of regional tertiary industry output value to regional GDP. (2) Regional human capital level. We use the number of local college students per 100 people (hcap). (3) the level of informatization. We use the total amount of local post and telecommunications services (comm) to represent the level of informatization. (4) Transportation conditions. We use regional highway mileage (total length of highway) to reflect it. (5) the level of science and technology. We use the number of local patents approved in that year to reflect. (6) Cultural factors. We use the consumption proportion (consu) of local urban residents except clothing, food and housing to reflect the consumption culture of a region (Jin, Tian Lin, 2004).