Abstract: It is obvious that optimal design of crop soil sampling locations would significantly decrease the number and costs of sampling by the premise of satisfying sampling precision, or insure the sample having the best representativeness by the limitation of sampling quantity. A multi-object optimization analysis and decision-making model of regional crop soil densities sampling distribution was put forward by researching and screening the various factors which influence the distribution of crop soil nutrients sampling destinies, and the spatial effect of different soil nutrients. The model's solving process is analyzed according to the improved genetic algorithm. At last, the model was analyzed and verified separately from two views for finding the best sampling densities distribution under the condition of assigned number, and insuring the minimum quantity of sampling densities by the premise of satisfying sampling precision and representativeness based on the farmland spatial data and screened environmental data in Baoying, Yangzhou. The results show that this model is suitable for solving soil nutrients uniform sampling planning problems under the condition of a large number of dispersive arable lands in the range of county area, which can provide a quantitative method for optimization analysis.