Abstract: For the geohazard susceptibility assessment in the regions where multi-hazards exist,the assessment system of evaluation factors should be established regarding the characteristics of various hazards.The creation of susceptibility zonation maps for different hazards is positive for the risk management and control and final decision-making.The mountainous area in the northern Tianjin City was taken as the study area in this study,and 2 sets of evaluation systems were determined in terms of the mechanisms of landslide,collapse and debris flows.The combined algorithms of statistic-artificial intelligence and slope unit were utilized to generate the susceptibility mappings for landslide-collapse and debris flows,respectively.The Receive Operating Characteristic (ROC) curve and potential geohazards identified by SBAS-InSAR were applied to verify the prediction accuracy,respectively.The results show that the susceptibility prediction performance based on historical hazard points and grid cell reaches up to 0.924 and 0.911,respectively.The number of potential collapse-landslide points and debris flows during the past 10 years recognized by SBAS-InSAR are 55 and 27,respectively.When the predictive susceptibility zonation is obtained by considering slope unit,the potential hazards located at high susceptibility level reach 92.7% (collapse-landslide) and 100% (debris flow),respectively.The results mentioned above indicate that incorporating the mechanism-data dual driven model into the geohazard susceptibility assessment is in line with reality and can provide a scientific reference for geohazard prevention and risk management in areas with multiple-hazards.