In order to improve the lag problem in the methanol catalytic combustion system that the temperature of the reforming chamber reaches the steady state due to the temperature shock effect, the fuzzy algorithm is used to optimize the traditional PI controller, and the P and I parameters are adjusted in real time based on prior knowledge, and the Matlab is passed software simulation calculation and design experiment compare the performance of the two control strategies. The simulation calculation and experiment results show that: in the rapid response stage of heating, the fuzzy PI controller has shorter response time and faster speed than the traditional PI controller; in the stable stage, compared with the traditional PI controller, the fuzzy PI controller produces more overshoot. Small, and can make the system temperature reach a steady state faster. It shows that the fuzzy PI algorithm can effectively improve the speed and stability of the system, and also improve system lag to a certain extent.