This paper aims at explaining the cross-border topological join queries of spatial fragments of the zone fragmentation data in distributed spatial database (DSDB), and the optimizing methods for such queries are proposed. First, the fragmentation and distribution of spatial data in a distributed environment are discussed, and the extra principles for spatial data fragmentation are put forward, including spatial clustering, non-partitioning on spatial objects, and maintaining logical seamless. Then, the fragment joins in zone fragmentation are classified into two categories: cross-border join and non-cross-border join; the topological relationships are also classified into two categories. Thus, the emphasis is put on the two types of cross-border topological joins. Two theorems for cross-border topological join optimization are proposed and proved. Based on the theorems, the optimizing rules for cross-border spatial topological join are given, including the removing rules and the transforming rules of fragment joins. Finally, tests are designed to compare three join strategies that include Na?ve join strategy, semi-join strategy and the proposed strategies. The results show that the proposed methods greatly improve the cross-border join optimizing efficiency. Therefore, the theorems and methods proposed in this work can be applied to the optimization of distributed cross-border spatial topological queries.