Abstract: Spatial data mining is to extract implicit spatial database,users are interested in knowledge.In this paper,the current clustering algorithm is not well taken into account the complexity of spatial data and the data link between the accuracy of clustering together is not high,designed a new algorithm-the space together based on information entropy Class of algorithm(ESCA algorithm),the algorithm complexity of spatial data priority and data link between the mechanisms by using ant colony optimization algorithm to improve the traditional shortcomings of cluster cluster number of uncertainties.Experimental results show that the algorithm is feasible and has higher accuracy.