As one of the spatial data mining technologies, DBSCAN algorithm is a densitybased clustering algorithm. Since it can find clusters with any forms from the spatial database, DBSCAN algorithm becomes more and more popular. The optimization principle and realization process of densitybased spatial clustering algorithm were studied in detail, and the existing problems of original DBSCAN algorithm were analyzed. By avoiding repeated searches of objects in the public domain, the computation of searches on the neighborhood of core object was reduced, and the time efficiency of the algorithm was improved. After analyzing the distribution of roadside stall business in rural areas, two key parameters, i.e., Eps and MinPts, of the algorithm and the searching zone of neighborhood of core object were determined. The experiment results showed that the time efficiency of optimized algorithm was improved by approximately 33.73%. Finally, the optimized algorithm was applied to the community grid management in rural areas. By data mining of the rural area grid management system, the most frequent regions were successfully identified for roadside stall business. Using this algorithm, the hot spots of problems in rural area management can be found out in time, which uncovered the common rules hidden behind the routine business. Hence, the corresponding management can be performed to a certain region, which can provide information and auxiliary decisions for rural area management.