Abstract: China has the third largest coal reserves and the first coal consumption in the world. China's coal resource areas are highly coincident with major grain producing areas or ecologically fragile areas. Under high-intensity mining, overburden strata move violently and surface damage is serious. The large-scale exploitation of underground coal resources will inevitably cause the destruction of surface attachments such as cultivated land, and disturbance of ecological environment in mining area, which will seriously affect the sustainable development of mining area. The surface damage induced by high-intensity mining includes static damage and dynamic damage, but the dynamic damage is more serious. Surface subsidence velocity can reflect the severity of surface movement and deformation. So, in this paper, we chose subsidence velocity as the dynamic deformation index to reveal the dynamic development law of surface damage induced by high-intensity mining in mining area. First, we took the field observed data of ground movement observation station set on the side of terminal line of 2407 working area as an example, and selected several periods of data in the full mining stage as samples. Through the analysis of mathematical statistics, we concluded that the subsidence velocity curve has the distribution characteristics of right skewness. Second, we established a physical similarity model test according to the mining and geological conditions of 2407 working area. Combined with this experiment, we found that the average of bedrock break angle was 56°, which revealed the formation mechanism of stepped crack. Under high-intensity mining, the bedrock was full-thickness fractured along the break angle of bedrock, the main failure type of loose layer was shear failure and the direction of failure was upward in the vertical direction. When the failure reached the surface, the stepped crack appeared on the ground surface. According to the analysis of the measured data of 2407 working area, it was found that overlying strata and surface had the periodic failure characteristics along the advancing direction of working area under high-intensity mining, the period of roof caving was the same as that of stepped crack, and the maximum subsidence velocity was located at stepped crack. The position of maximum subsidence velocity can be determined by the break angle of bedrock in real time. Then, we selected the lognormal density function with right skewness and normality, and combined with the real-time location distribution characteristics of maximum surface subsidence velocity to construct the dynamic prediction model of subsidence velocity. The reliability of the dynamic prediction model was analyzed by selecting the field measured data of subsidence velocity from adjacent high-intensity mining working face of Bulian Tower coal mine. The results showed that the correlation coefficient between predicted data and measured data was over 0.9, and the ratio of standard deviation to measured maximum subsidence velocity was less than7.0%. It showed the high reliability of the model. This research can provide guidance for dynamic prediction of mining subsidence in similar mining conditions.