In order to ensure the rationality and environmental friendliness of urban land use, a multi-objective spatial optimal allocation method of urban land use was proposed. Taking advantage of the advantage of shuffled frog leaping algorithm (SFLA) in solving multi-objective problems, a spatial optimal allocation model of urban multi-objective land use was constructed based on SFLA, with the minimum distance between the newly developed land and the developed land and the minimum incompatibility of environmental factors between urban land units as the constraint conditions of the objective function. Taking the urban land grid as the basic unit of operation, the improved shuffled frog leaping algorithm, such as head and tail exclusion grouping, intelligent learning and mutation operator, is introduced to obtain the optimal solution of the urban multi-objective land use spatial optimization allocation model. The experimental results show that the environmental compatibility is almost above 0.5, and most of them are close to 1 after the optimal allocation of urban land. It can better realize the optimal allocation of urban multi-objective land use space with high efficiency, and the conservation and environmental compatibility of land resources after optimal allocation are also better.