Abstract: Quantification of geological disasters risk at urban scale under different rainfall intensity is difficult points during geological disasters risk assessment. Taking the landslide chain disaster in cluster shallow of red bed area as a case study, the authors in this paper explored a new risk quantitative evaluation method of geological disaster at town scale to provide a foundation for the risk assessment of urban geological disaster. By querying the statistical parameters of rainfall under different rainfall frequencies in the Mihe area in Xide Town, the authors census the data of the national rainfall station and the distribution characteristics of rainfall patterns of 18 geological disasters in the past 50 years. Through the statistics of soil thickness, extraction of vegetation coverage and terrain data processing, the risk of geological disasters in cities and towns was calculated based on STEM TRAMM numerical calculation method and rainfall distribution curve, and the risk assessment map of geological disasters in the study area was drawn. Through the comparison of remote sensing interpretation data, ground survey data and disaster database data with numerical calculation results, the authors find that it has good adaptability, convenience and science of the application of rainfall characteristic statistics and STEM TRAMM numerical calculation method on fine evaluation of urban geological disaster risk in red layer area, which can provide some ideas for the hazard assessment of urban geological disasters under different disaster pregnant background.