Data model in information space is the basis of effectively managing heterogeneous and interrelated data sources. However, traditional data modeling approaches fall short when representing context-dependent heterogeneous information and complex semantic associations, and when supporting semantic association reasoning. To overcome these drawbacks, in this paper, we propose a semi-structured graph model named context-aware and complex semantic association network model. In particular, we introduce the notion of context-aware interpreted object which encapsulates heterogeneous information about the underlying data and context information. We descript model complex semantic associations in terms of a set of constraint components such as order constraints, aggregation constraints, and attribute constraints. Formally, we define the semantic association rule in favor of semantic association reasoning. Also, we propose a snapshot generation algorithm for our model and further illustrate the snapshot of the data graph according to our model.