The effect of urban planning is related to the happiness of the people’s life, and it is necessary to carefully process and effectively integrate the geographic information involved in it, and endeavor to improve the integration degree of each region of the city. In this paper, the relative coordinate positions of geographic images before and after geometric correction are calculated to find suitable corrective control points and improve the accuracy of mapping information. The identification framework and multi-class function of D-S evidence information fusion theory are utilized to effectively identify different geographic images and complete the classification of urban elements. Introducing the spatio-temporal data model for urban geographic element objects to complete the perceptual fusion of geographic information and optimize urban planning. The D-S evidence information fusion theory model achieves the highest classification accuracy of 97.61% for the seven functional areas, and the time required is no more than 1.2 s. The spatio-temporal data model fusion of geographic information reduces the degree of overlap of the planning of various areas of the city to less than 0.4%, and enhances the rationality of urban planning.
