Abstract: In order to further improve the accuracy and reliability of mining subsidence prediction,a Boltzmann function prediction model coupled with the correction of inflection point offset and its parameter inversion method were carried out. Based on Boltzmann function,a Boltzmann function prediction model coupled with the correction of inflection point offset was constructed by discussing the correction method of inflection point offset. In order to obtain the model parameters accurately and reliably,the FWA was introduced,a parameter inversion method of Boltzmann function prediction model coupled with the correction of inflection point offset (BFWA-MSPP) was proposed. The results of the simulation experiment show that the average relative error of thesolved parameters by BFWA-MSPP was approximately 2.1%,with 2.3 of average RMSE of the solved parameters,and the error of the fitting subsidence and horizontal movement were between -11.6~5.7 mm,with ±6.1 mmof average RMSE. In addition,after adding a large number of random errors and gross errors to the observations,BFWA-MSPP can inverted model parameters more accurately,which weaken the influence of observation errors to a certain extent. The BFWA-MSPP and PIM-MSPP were applied to the inversion prediction parametersof the 1414(1) working face of Guqiao Coal Mine respectively. The RMSE of fitting the subsidence andhorizontal movement were ±83 mm and ±108 mm,respectively,and the result of the solved parameters by BFWAMSPP was as follow:q=1.07,tan β=1.91,b=0.41,θ=88.32°,S₁=-18.89 m,S₂=14.74 m,S₃=48.89 m,S₄=64.42 m.