Tropical cyclone (TC) is considered as the biggest weather event that can cause severe damage in power distribution grid. The predicting of the TC induced damage plays significant role to prepare the mitigation action or to plan a rapid restoration process. In this manner, this present study focused on understanding and predicting the damages associated with TCs pertaining to the number of affected customers considering TC track. Three types of historical data such as TC data, weather, and distribution grid damages are used in this study. Then, describes and analyzes the historical data and presents the characteristic of damages induced by TCs. The impact of TC track shows by directly comparing the damages incurred in TC landfall regions. Besides, the statistical analysis is presented to validate the track dependence of the damages, and the random forest analysis is used to identify the priorities of TC and weather variables. Lastly, a multistage damage intensity prediction method is proposed utilizing track information as well as the priorities of variables. The proposed method approximately predicts the intensity of damages caused by TC over a considerably wider geographical area.