Abstract: To explore a new numerical forecast way for thunderstorm, a high-resolution mesoscale synoptic-dynamics parameterization and electrification-lightning unified forecast model (GRAPES_LM) (Global/Regional Assimilation and Prediction System_Lightning Model) has been developed, based on the GRAPES_Meso system. The addition of a range of needed components, including the equation sets for predicting the charge density of the different water substances, and charging and discharging process parameterizations which has been splited from the traditional thunderlightning model, has enriched the meso-scale system with the parameterized synoptic dynamics process, electric process, and physical processes. And thus a highresolution GRAPES_meso lightning forecast model aimed to predict lightning thunderstorm has been established. The new GRAPES_meso lightning forecast model has been incorporated the various parameterized components for the sophisticated microphysical-processes,including the dual-parameter-convective-cloud parameterization sheme developed by the Chinese Academy of Meteorological Sciences, the three charging parameterization schemes for the noninductive rebounding graupel (hail)ice (snow) collisions charging, the inductive graupel (hail)cloud droplets collision charging, and the graupel-large cloud droplets collisions caused secondary ice crystal charging, as well as the latest random bi-directional lightning parameterization scheme. To validate and evaluate the predictability for its lightning forecasting, the model was used to simulate the CCOPE (Cooperative Convective Precipitation Experiment) case of 19 July 1981. The results show that the high-resolution GRAPES_LM simulated successfully the thunderstorm case occurred on 19 July 1981, and the spatial and temporal development of the dynamic, microphysical, and electrification processes agree with the observed results quite well. This indicates that the lightning numerical forecast is feasible and it provides a new tool for lightning study. 