A mesoscale convective progress occurred in Guangdong Province on 16 June 2017, and the system was simulated by WRF model. This article analyzes the impact of the combined assimilation of the lightning and conventional surface observation data on the simulation of mesoscale convective system compared with the single assimilation of one kind of data. The lightning data were continuously assimilated into the model through the WRF-FDDA system with a lightning accumulation window of 15 min, while the conventional observation data were assimilated into the model by the WRFDA-3DVAR system with one hour interval. The results show that the introduction of lightning data in the joint assimilation experiment has improved the accuracy of updrafts, cold pools, and gust fronts in the background fields relative to the assimilation of conventional surface observations only. The introduction of conventional surface observations has reduced the background field errors in temperature, water vapor, and wind fields over a larger area, suppressed the spurious convection in some areas, and overall improved the simulation accuracy of the convective system. The results of prediction skill score show that the combined assimilation of the two kinds of data can also improve the prediction skill score of the assimilation period and the forecast period to some extent.