Abstract: According to deficiency of the Knothe time function in describing the dynamic subsidence prediction process in the mining subsidence area, a new dynamic subsidence model is proposed, three-parameter Knothe time function. The initial settlement speed parameter b1, the time power index parameter b2 and the curve shape parameter b3 are added to the model. The parameters solution is based on the particle swarm optimization (PSO) algorithm. The measured data proves that the dynamic subsidence prediction model of the mining area based on the improved Knothe time function can reflect the dynamic process of surface subsidence. The maximum error between the measured value and the predicted value of the strike line is 5.02 cm, the minimum error is 0.1 mm, and the average error is 1.19 cm in each observation period. The accuracy is very reliable and meets the needs of mining work.