System dynamic model was established based on the need of coldside temperature control in the power measuring system of radioisotope heat source. On the basis of the openloop simulation analysis

three control algorithms

namely

conventional proportion integration differentiation (PID) control

fuzzy control and support vector machine(SVM) intelligent predictive control

were designed respectively

with which the feasibility of temperature control of the measuring system could be verified. Results suggest that

the power measuring system presents high robustness and stability by combing the excellent nonlinear function approximation ability of SVM and the predictive control method to fit the nonlinear system model

and by correcting the data based on the feedback of the SVM predictions.