Abstract: The examination of reconstruction for missing data and homogeneity for summer precipitation and temperature in China was completed. The experimental study is on the reconstruction for missing data for the period of 1~5 years using stepwise regression and stepwise discrimination. It is found that the stepwise discrimination is better than the stepwise regression in the validity of the reconstruction. The experimental studies about different methods of discriminate group show that the discrimination based on Gamma distribution is better with a good validity. The discrimination selected nearly stations have a good validity than the stepwise discrimination with F value. The experimental results also show that the reconstruction for missing with the little error in the six areas in China on reconstruction using discrimination selected nearly stations.