Abstract: An improved fuzzy control method based on adaptive neuro-fuzzy inference system（ANFIS）is proposed in order to address the issue that the operation environment of solar-air source heat pump heating systems is complex and variable,and the design of fuzzy controller heavily depends on artificial experience. This technique establishes the off-working condition ANFIS model of the combined heating system,generates the fuzzy rule base that adapts to the system performance,and passes it to the fuzzy controller for execution. It makes use of the benefits of ANFIS in complex system modeling,along with the temperature response characteristics of the heating system and the actual operation data. According to the findings of the Matlab/Simulink simulation,the control precision and response speed of this approach are greater by 17.65% and 36.4%,respectively,when compared to a single fuzzy control.