Huge amounts of spatial data have been accumulated by government agencies and other organizations for various kinds of applications, such as land information management, asset and facility management, resource management, and environment management. These data are typically collected, stored and used by different custodians for very specific purposes. The full benefits of using spatial data can be achieved by combining the data from different sources covering a common region. It is, however, a great challenge to use the spatial data collected and organized for one purpose for other purposes. In comparison with integration of traditional relational databases, there are a number of unique problems faced by spatial database integration, such as data resolution and precision, data models and representation, and spatial indexes. In this paper, we will identify these issues and discuss possible solutions.