The mean generation function, standard normal uniformity test, correlation analysis and other methods are being used to interpolate, check, correct, and analyse the long-term summer precipitation data of 96 observation stations in eastern China for the period from 1931 to 2020. The results show that: (1) The overall trends and extreme values of the summer precipitation data of all stations from 1931 to 2020, fitted by the mean generating function, are in good agreement with the observed values. The consistency rate of six stations without missing data between the observed and fitted values reaches 86.1%, which can meet the needs of the interpolation work. (2) As for the 1931-1950 period with missing records and the 1951-2020 period without missing records, summer precipitation is being tested for differences using the mean and variance statistics. Only eight observation stations have showing significant differences. (3) There are 13 non-homogeneity stations of the interpolated data in the homogeneous testing and checking of the summer precipitation data from 1931 to 2020. (4) The spatial distribution similarity analysis is being conducted between the corrected station data and the grid data of the CRU_TS4.05 database. The spatial correlation coefficients of the two sets of data during 1931-1950, 1951-2020 and 1931-2020 have reached 0.90, 0.92 and 0.92, respectively. The spatial distribution is consistent, and the corrected station data show certain reliability.