How to judge the achievement of training objectives, analyze whether training will bring knowledge change and ability improvement to trainees, and finally bring benefits to enterprises and society, and training effect evaluation management is increasingly important in modern enterprises. Multi-mode training effect evaluation kirkpatrick model is the most widely used model at home and abroad. Professor Donald from the University of Wisconsin? Kirkpatrick proposed in 1959 that he divided the training effect into four levels according to the depth and difficulty of evaluation: reaction level, learning level, behavior level and result level. The reaction layer is the trainee's reaction and evaluation to the training project, and it is the lowest layer in the evaluation of enterprise training effect. Including the satisfaction of trainers, training management process, testing process, course materials and course structure. Behavior layer Behavior layer is to measure the degree of transformation of skills and knowledge learned in training programs and whether the work behavior of students has been improved. In this regard, the behavior changes before and after training can be evaluated by superiors, subordinates, colleagues and trainees themselves. Kaufman extended kirkpatrick's four-level model. He believes that obtaining various resources before training is very important for the success of training, so this level of evaluation should be added to the model. He believes that the effect of training is not only beneficial to the organization, but also will eventually affect the environment in which the organization is located, thus bringing benefits to the organization. So he added a fifth level, which is to evaluate the reaction of society and customers. Stage evaluation model CSE evaluation model was put forward by the evaluation research center of the University of California. The evaluation model is characterized by evaluating the whole training process in stages, so it is not limited by time and conditions and is easy to operate. Compared with other evaluation models, its advantages are mainly reflected in the combination of stage, comprehensiveness and whole process evaluation. Because it evaluates the occurrence and development of the whole training in stages, it can effectively obtain the reliable information of each stage and link in the training process, and constantly control, adjust and improve the training work. The CIPP model is similar to CIRO, which holds that evaluation must be carried out from four aspects: situation, input, process and result. The problems existing in the training evaluation of enterprises in China, the high-level enterprises and human resources workers generally pay insufficient attention to the training evaluation. A questionnaire survey for enterprise leaders and trainers shows that only 10%- 15% of enterprise training has been evaluated, and 60% of respondents think that training evaluation is not important. The evaluation of training effect only stays at the primary level of evaluation. At present, the most widely used evaluation method of enterprise training effect in China is still the four-level training evaluation model of American scholar kirkpatrick, but most of them are only evaluated in two stages: reaction layer and learning layer. As for the evaluation of behavior level and result level, it is time-consuming, laborious and expensive to implement, and the cost-benefit of training is difficult to quantify, which is relatively rare in enterprises. The tools and means used in training evaluation are single. At present, the evaluation tools adopted by most enterprises are questionnaire survey and after-class evaluation. Other evaluation tools, such as interview, skill practice, on-the-job training, behavior observation and comparison group, are rarely used, which may lead to the evaluation tools not effectively reflecting the contents of training programs. Training evaluation lacks quantitative indicators. Scientific and effective evaluation should be a combination of quantitative and qualitative to ensure the scientific accuracy of evaluation. Lack of quantitative indicators is a common problem in many enterprises at present. The main reason is that the training evaluation system itself is not perfect, and it is difficult to obtain variables and the long period is also the difficulty of this work, which makes the training evaluation lose its scientificity.