one of the key activities in service-oriented solution development is the identification of services according to a set of predefined design principles. Existing Top-Down technique is normal way for service identification and some relative further research are based on it. They all paid more attention to how to automatic decompose business and take bottom/leaf actives as a candidate service while ignore service character, such as loose couple and high cohesive. The purpose of this paper is to provide a way to optimize business process for Service Identification. The study was composed of four majors parts: (1) a directed graph model is built to describe business process simply by employing Graph Theory, (2) a five steps identification produce is given and some formulas are present to, (3) Based on DBSCAN (Density-Based Spatial Clustering of Applications with Noise), a network topology Clustering algorithm was given and (4) an example has been conducted to show the applicability.