Abstract: Spatial clustering is an important tool for spatial data mining and spatial analysis. It is important for the clustering results in many applications to meet the requirement that spatial objects in the same cluster are similar in both the spatial and the attribute domains. To solve the problems of existing methods based on Delaunay triangulation, this paper proposes a spatial clustering method considering uneven distribution of non-spatial attributes. The proposed algorithm involves two main steps:the first is to construct spatial proximity relationships, and the second is to cluster spatial objects with similar attributes. Delaunay triangulation with edge length constraints is first employed to construct spatial proximity relationships among objects. To obtain satisfied results in spatial clustering with attribute similarity, the breadth first traverse (BFT) clustering algorithm is used in accordance with variation rate of attribute to adapt to the local change information of attribute distribution. The performance of the proposed algorithm was evaluated experimentally through comparison with one of the leading state-of-the-art alternatives:multi-constraints algorithm. The results show that our method outperforms the comparative algorithm as attributes are unevenly distributed in space, and provides a quantitative research method in city commercial center extraction. The effectiveness and practicability of the proposed algorithm illustrates three advantages of our algorithm:i) it can reflect the tendency of the entity attribute in the spatial distribution; ii) it can meet the requirement that attributes are unevenly distributed in space; iii) it can discovery clusters with arbitrary shape and is robust to outliers.