Geological condition and rainfall are two most principal conditions inducing landslides in the Chongqing region. By analyzing the forming conditions of rainfall-induced landslides, a new method for spatio-temporal prediction of rainfall-induced landslide is brought forward on the basis of grading and overlapping geological condition and rainfall factor in this paper. At first, semi-quantitative assessment and grading for the geological condition of a certain area or slope can be carried out with the multi-factor interactive matrix. Then the severity of rainfall in that area is grading according to the maximum daily rainfall and the total rainfall in a rainfall course. Finally, the “landslide probability judgement factor” can be worked out through grading and overlapping “geological condition influenc- ing factor” and “rainfall influencing factor”, by which the landslide can be graded into 4 grades, they are landslide extremely easily happening, landslide easily happening, landslide difficultly happening and landslide hardly ever happening respectively. More accurate spatio-temporal prediction of rain- fall-induced landslides can come true on the ground of detailed geological survey of some dangerous slopes in an area and more precise weather forecast. Finally, the reliability and feasibility of carrying out the spatio-temporal prediction of rainfall-induced landslides with the method of “two factors” grading and overlapping are validated by the example of Jipazi landslide. Geological condition and rainfall are two most principal conditions inducing landslides in the Chongqing region. By analyzing the forming conditions of rainfall-induced landslides, a new method for spatio-temporal prediction of rainfall-induced landslide is brought forward on the basis of grading and overlapping geological condition and rainfall factor in this paper. At first, semi-quantitative assessment and grading for the geological condition of a certain area or slope can be carried out with the multi-factor interactive matrix. Then the severity of rainfall in that area is grading according to the maximum daily rainfall and the total rainfall in a rainfall course. Finally, the “landslide probability judgment factor” can be worked out through grading and overlapping “geological condition influencing factor” and “rainfall influencing factor”, by which the landslide can be graded into 4 grades, they are landslide extremely easily happening, landslide easily happening, l andslide difficultly happening and landslide hardly ever happening respectively. More accurate spatio-temporal prediction of rain- fall-induced landslides can come true on the ground of detailed geological survey of some dangerous slopes in an area and more precise weather forecast. Finally, the reliability and feasibility of carrying out the spatio-temporal prediction of rainfall-induced landslides with the method of “two factors” grading and overlapping are validated by the example of Jipazi landslide.