The automatic control system of heating ventilation and air conditioning (HVAC) is the most important system in building automation. From the point of view of saving energy and improving indoor environmental quality, building automation plays an increasingly important role in HVAC. Therefore, it is of great significance to adopt effective air conditioning methods for energy management and air quality control of intelligent buildings. At present, variable air volume (VAV) air conditioning system has gradually become the mainstream of air conditioning systems at home and abroad because of its huge energy-saving potential. In order to better apply VAV air conditioning to practical projects, this paper mainly does the following work: firstly, by analyzing the direct reasons for the low humidity control accuracy of VAV air conditioning, through numerical calculation and analysis, a VAV air conditioning system with variable dew point control is established, and the effectiveness and feasibility of this method are analyzed theoretically; Elman neural network model for dew point temperature prediction is established. Experiments show that the prediction method can well predict the optimal dew point temperature of VAV air conditioning system in the future, which provides the possibility for variable dew point temperature control of VAV air conditioning system. In order to truly solve the control of variable dew point temperature and improve the accuracy of temperature and humidity control from the control point of view, this paper establishes the transfer function between water valve and air supply temperature by using the method of system identification, which verifies the feasibility of variable dew point temperature control and the effectiveness of improving the accuracy of temperature and humidity control. Through experiments, this paper once again proves that variable dew point temperature control can not only improve the accuracy of indoor temperature and humidity control, but also verify that this method can save energy and realize the rational use of energy from the analysis point of view. Finally, in order to better control the dew point temperature in VAV air conditioning, a PID control algorithm based on BP neural network is proposed. The simulation experiment further shows that this control algorithm has better control effect than PID control algorithm.