Precision temperature control systems are critical in semiconductor manufacturing to maintain the stability of the ambient temperature. However, determining the optimal control parameters becomes difficult due to the characteristics of the system such as time lag, nonlinearity and large inertia. In addition, when the system state changes, the control parameters need to be recalibrated to adapt to the new operating conditions. The existing controller parameter tuning methods are mainly divided into engineering tuning method and optimization algorithm. The former is easy to use and convenient to calculate, but it is only applicable to the scenarios with low control accuracy requirements, and it depends on the experience of the operator, which is time-consuming and labor-intensive as it needs to be adjusted repeatedly. Although the latter performs better in terms of parameter optimization efficiency and control accuracy, it is difficult to ensure the stability of the system, and there is a risk of falling into the local optimum. Aiming at the above problems, this paper proposes a controller parameter self-tuning method based on stability region performance optimization. The method firstly obtains the system model through model identification, and then applies the stability region locus method to determine the stability region of the controller parameters on this basis. In order to evaluate the set value tracking ability and disturbance resistance of the system, corresponding performance indexes are proposed. The stability region is divided into grids and the performance indexes corresponding to the grid point parameters are calculated to ensure that the calibration results are globally optimal. In addition, this method uses actual system data as model perturbation inputs for performance evaluation, which makes the tuning results more reliable. Experimental validation shows that the controller parameters obtained based on the method of this paper have significant advantages in system control accuracy compared with manual parameter tuning.