Abstract: The overburden subsidence of coal seam mining is ahead of the surface subsidence and can be used as an early warning indicator of surface stability. This paper studies the problem of overburden subsidence prediction caused by coal seam mining, uses distributed optical fiber sensing technology, makes a similar material test model for mining overburden in Daliuta Coal Mine, and study the deformation characteristics of overburden caused by coal seam mining. Based on grey theory and Knothe time function, the paper further establishes a prediction model of mining overburden settlement based on measured strain, and discusses the accuracy of the prediction model. It achieves the following research results. Overburden strain distribution based on distributed optical fiber sensing technology accurately reflects development characteristics of falling zone and water-conducting fracture zone. The prediction accuracy of overburden settlement based on GM(1, 1)model and metabolic model is better but limited to short-term prediction. According to the calculation method of reverse parameters when the relative residual maximum value and the sum of relative residual take the minimum value, the settlement parameters c value and τ value are solved, which improves the accuracy of the Knothe time function prediction. The paper also establishes the expression of dynamic subsidence prediction model of overlying strata in Daliuta coal seam, which is suitable for the early stage and later stage prediction of overlying strata subsidence.