煤炭资源的大规模开采，会引起或诱发煤矿区地面沉陷、塌陷和山地开裂等地质灾害，不仅严重危害了当地的自然环境，也给当地人民群众及矿区生产造成巨大的经济损失，对开采沉陷的预测，不仅有助于采取适当的措施加以控制，同时对于后期矿区环境的恢复与整治也非常需要。目前地表移动预计的方法较多，但由于对开采沉陷机理及地表变形破坏的认识、参数取值等众多因素的影响，预计结果与实际仍有较大误差。 本文通过分析受采动影响的上覆岩层的变形特征及主要控制因素，基于分层传递原理，运用关键层理论及概率积分方法，尝试建立了一种空间上的地表下沉移动变形模型，并在此基础上推导出了倾斜、曲率、水平移动和水平变形公式。之后结合Knothe时间函数，通过叠加原理，推导出了开采引起地表移动变形的时空预计模型。并给出了该预计模型中参数的确定方法。对于有限开采条件下，在计算砌体梁断块长度时，通过引入初次来压步距与工作面长度的关系，增加了模型参数的合理性和准确性。 结合陕西省大佛寺煤矿区进行实例验证，在计算过程中运用matlab软件对形成的模型求解，得到的理论预测结果与实测资料吻合的较好。与常规概率积分法的预计结果进行对比，结果表明：利用本模型预计下沉值最大相对误差为32.54％，相对误差平均为10.36％。采用概率积分法的预计下沉值最大相对误差为99.90％，相对误差平均为41.14％。说明构建的模型是可行的，且效果要好于常规概率积分法预计结果。然后利用时空模型对大佛寺煤矿40301首采工作面推进490m、581m、672m、700m及停采后0.5a、1a、2a的下沉进行了预计分析，证明了构建的时空模型的可行性。

Large scale mining of underground coal resources may cause or bring out hazards of geological environment such as ground subsidence, fault, fracture of the ground, not only seriously endangered the local natural environment, but also caused huge economic losses to the local people and mining production.The prediction of mining subsidence is helpful to take proper measures to control surface movement and subsidence , and is very important to recover and improve the post mining environment in the area.At present, there are many methods of analyze the problems of ground displacement, but there is still large errors between estimated results and the measured value due to insufficient knowledge of mining subsidence mechanism and the surface deformation caused by Underground mining, determination of the parameters and so on. Based on hierarchical transfer principle, this article established a new type of ground subsidence and deformation model by using the key strata theory and probability integral method after comprehensively analyzed the deformation characteristics of the overlying strata and the main control factors resulting from mining subsidence,and deduced the formula of slant, curvature, level movement and level distortion based on the model. After that, considering the Knothe time functions, the subsidence prediction model of time and space due to underground mining was deduced out,and the methods of the various prediction parameters in this model were also found. Under the situation of sub-critical extraction, the article introduced the relationship between the working face length and the first weighting step distance to calculate the length of the stacked layer fault block, making the model parameters be more reasonable and accurate. Combined with practical examples of Dafosi coal mine of shaanxi province to validate the new prediction model, and used software matlab to solve the formed model in the process of calculation.The results show that the most relative error of the model built in this work is 32.54%, the average relative error is 10.36％. the most relative error of conventional probability integral method is 99.90％, the average relative error is 41.14％. It proved that the model built in this work is feasible, and the effect of the model is better than the expected results of conventional probability integral method. Then apply the space-time model built in this work to the first coal face 40301 in Dafosi coal mine to analyze the mining subsidence as the working face proceed to 490m, 581m, 672m, 700m and 100 days after stopping mining, by this way, the feasibility of the constructing space-time model is being proved.