Abstract: A fuzzy neural network temperature-control system based on thermoelectric cooler (TEC) was proposed and designed. The hardware of this system consisted of temperature sensor, microcontroller, thermoelectric cooler and its driver. The software of this system adopted fuzzy control algorithm and neural network algorithm. The output of the fuzzy neural network (FNN) was loaded into the MCU to control the PWM waveform generator, which could generate suitable waveform to drive the TEC. Simulations of the system show that the proposed FNN temperature-control system had better performance than the conventional PID controller.