Abstract: In view of unreasonable evaluation factors and limitations of regional evaluation results of the current mining-induced building damage, a method is proposed for assessing building damage based on surface sensitive deformation and housing factors. Mining-induced building damage evaluation is studied by means of theoretical analysis, numerical simulation and field investigation. Firstly, numerical simulation method is used to reveal the influence rule of surface deformation, building structure type, the angle between the main deformation direction and the long axis of the building on the building damage. The results show that ① when the deformation along the long axis of the building increases, the degree of damage to the building increases. It is obtained the critical value of surface deformation for the accelerated damage of one-story bungalow building is curvature K=0.6 mm/m² and horizontal deformation ε=6 mm/m. ② The damage degree of buildings with different structures is different, in which the buildings with ring beams or structural columns have a strong deformation resistance. This study has determined the critical surface deformation value of the damage degree of the two-story brick-concrete structure with ring beams and structural columns. ③ The building is more sensitive to the surface deformation in the direction of its long axis, affecting the damage is large. However, the deformation of the perpendicular direction to the long axis of the building is not very sensitive, and the damage degree at the oblique intersection of principal deformation direction and the long axis of houses is in between the two. Secondly, the unascertained measure model of mining-induced building damage is established, and a total of 7 evaluation indicators are selected based on surface sensitive deformation factors, principal deformation direction, and house factors, then the weight is assigned to each indicator using the Analytic Hierarchy Process (AHP) according to its importance. Finally, it is applied to the building damage cases of three villages in the Fengfeng Mining Area and the Xuzhou Mining Area, and the application shows that the prediction accuracy is above 80% in the evaluation of more than 140 buildings in 50 households selected. The prediction accuracy is higher than the traditional method, and the prediction effect is good, which makes up for the shortcomings of traditional evaluation methods and verifies the reliability of the evaluation index selection and the evaluation model.