Abstract: Semantic similarity measurement is the key technology to realize the integration and fusion of multi-source vector spatial data. Firstly, this paper analyzes and describes the semantic information of geographical entities from the perspective of vector spatial data representation, and proposes a semantic similarity measurement model based on multi-feature constraints.Secondly, the model uses the classification relation of geographical elements to control the extraction of the target entity set. On the basis of constructing the corresponding relationship of semantic features among entities, the concept of attribute feature entropy is introduced to calculate the weight values of different features, and then the overall semantic similarity of geographical entities is measured by synthesizing the multi-feature similarity. Finally, the model is applied to road entity matching. The road matching is realized by calculating the semantic similarity between entities. Meanwhile, the validity of the model is verified. The experimental results show that the semantic similarity measurement model based on multi-feature constraints can reasonably calculate the semantic similarity of geographical entities and improve the efficiency of semantic matching of geographical entities.