This article proposes the hybrid method to inverse the equivalent electric charge of thunder cloud based on the data of multi-station atmospheric electric field. Firstly, the method combines the genetic algorithm (GA) and Newton method through the mosaic hybrid structure. In addition, the thunder cloud equivalent charge is inversed based on the forward modeling results by giving the parameters of the thunder cloud charge structure. Then an ideal model is built to examine the performance compared to the nonlinear least squares method. Finally, a typical thunderstorms process in Nanjing is analyzed by Genetic-Newton algorithm with the help of weather radar. The results show the proposed method has the strong global searching capability so that the problem of initial value selection can be solved effectively, as well as gets the better inversion results. Furthermore, the mosaic hybrid structure can absorb the advantages of two algorithms better, and the inversion position is consistent with the strongest radar echo. The inversion results find the upper negative charge is small and can be ignored, which means the triple-polarity charge structure is relatively scientific, which could give some references to the research like lightning forecasting, location tracking.