Abstract: When Ray Tracing Algorithm is used to simulate the possible sunshine duration under undulating terrain， the different search radius of the terrain shading condition affects the accuracy and efficiency of the calculation of the possible sunshine duration. In the report， based on DEM data， aimed at the unique orbital motion characteristics of Mercury and the slow change of the solar elevation angle with the motion of Mercury， the search radius of different solar altitude angles and the change of average possible sunshine duration were investigated. At the same time， the BP neural network was constructed with five factors affecting the search radius and the searchable time as input variables and the search radius and the possible sunshine duration as output variables. The model passed the test， and the significance of the influence of the five influencing factors and the optimal search radius from high to low was： solar elevation angle > elevation standard deviation > average of terrain width > terrain relief > average of surface roughness， and the significance of the influence of average possible sunshine duration from high to low was： solar elevation angle > elevation standard deviation > average of surface roughness > average of terrain width > terrain relief. The model can be used as a reference for calculating the searching radius and average possible sunshine duration of Mercury.