Thermoelectric cooler has dynamic thermoelectric performance under complex environment. A linear dynamic model of the thermoelectric cooler is derived using small-signal linearization method. It shows that the dynamic model of the thermoelectric cooler has one zero and two poles. The dynamic model of the thermoelectric cooler is showed to vary with operating conditions. Based of average linear dynamic model of a thermoelectric cooler, a temperature control system is designed for the cold end temperature of the thermoelectric cooler using fuzzy-PID algorithm. The step response tests show that the controller has satisfying dynamic and static performance. In room temperature environment, the response time for cooling down 18°C is around 340s and the steady error is very small. The cold-end temperature can be maintained at the setting value within 0.1°C. Experiment results also show that the setting temperature can reach to 23°C below the environment temperature, and the setting value is smaller and the step response time is longer.