<p>Land subsidence caused by man-made underground coal mining has always been a common engineering geologic problem. At present, there are many researches on the amount and extent of land subsidence caused by coal mining. However, the loose layer drainage disturbed by coal mining can also cause land subsidence, the extent of loose layer drainage is larger than that caused by coal mining, which increases the deformation risk of structures located at the boundary of a mining area. There are few researches on the coupling effect of coal mining and loose layer drainage especially in Shanxi Province, China, where groundwater is invaluable. This paper proposed a practicable land subsidence prediction model considering the coupling effect based on Boltzmann function. The parameters of the proposed model were obtained by the dynamic step fruit fly optimization algorithm according to measured subsidence data of working panel 1309 in Shanxi. The root mean square error (RMSE) is 20.9&#xa0;mm, and the relative RMSE is 1.2%. Compared with the same evaluation indexes using the model based on the probability integral method, the accuracies are both increased more than 50%. Land subsidence along the national coal railway above the boundary of working panel 1309 was also predicted. The predicted subsidence was compared with fluid–solid coupling simulation results, and the predicted results are satisfactory. This model is expected to have good applicability to the land subsidence of coal mining considering loose layer drainage without too many geomechanical parameters which are difficult and cost to obtain. It can accurately predict land subsidence caused by coal mining under the condition where the loose layer contains water. The prediction result can reconcile the conflict between underground coal mining and the protection of ground structures within the mining-affected area.</p>