Abstract: For the problems of great investment, short service life, big measurement errors, etc. in the zirconia sensors for the power plant, the support vector regression algorithm is applied to dealing with the data collected in the distributed control systems of power plant. Kernel function k(x_k;x_l) is used to realize mapping from low dimension to high dimension instead of the inner product process in the nonlinear problems. So the low dimension nonlinear regression problem is transformed into the high dimension linear regression problem. The selection methods of auxiliary variables, kernel functions, kinds of parameters are given. The feasibility of this method is verified by a simulation experiment.