Abstract: Xuzhou is an industrial city which extracts the largest amount of groundwater for water supply in Jiangsu province. Uncontrolled extraction from the fracture-karst aquifer has caused many environmental geologic problems, such as regional cones of depression, specific capacity decreasing in wells, groundwater quality deterioration, and karst collapse. With the development of the urban size and local economy, the water demand for the city will increase sharply. Therefore, rational utilization of groundwater resources is necessary and urgent for the Xuzhou City. 　　The research area covers five well-field districts. In this study, the five well fields are simulated as a unity. First, porous media model is applied to modeling the fracture-karst flow based upon a simplified conceptual model maintaining the key features of the research area. Then, Finite Element Method (FEM) is used to solve the corresponding mathematical model. The numerical simulation strictly follows the national standards of P. R. C. (GB/T 14497-93) Requirements for the Work of Groundwater Resources Management Model to guarantee the accuracy of study and increases the reliability. The results of modeling demonstrated that the theory of porous media model is suitable for the regions of fracture-karst in North China. In the previous studies, the average rainfalls or the frequency combination of precipitation were used to evaluate groundwater resources, which usually yield a result too conservative to exactly appraise groundwater resources. Therefore, in this study we employed a nonlinear time series model for precipitation in the research area, and apply the model to forecast the rainfalls in the cause of evaluating groundwater resources. We believe this measure could make a more accurate prediction. 　　According to the results of this research, which are compared with the actual statistical data of the current exploitation, the current extraction conditions of Xuzhou City are irrational. 　　Such irrationality lies not in that the current exploitation is over allowable groundwater yield on the whole, but that there is no scientific distribution of exploitation in each well field. 　　If the administrator adjust the existing exploitation conditions in accordance with our research results, the environmental geologic problems will be gradually resolved, and the present contradiction between water supply and water demand in the research area will not be further sharpened. This paper shall be of profound significance to solve the contradiction between water supply and water demand in the regions similar to the research area. Of course, the efficiency using the method of combining the time series model and FEM in the course of long-term groundwater resource evaluation should be verified in the further work.