These days the ability to prove an individual identity is crucial in social, economic and legal aspects of life. Identity resolution is the process of semantic reconciliation that determines whether a single identity is the same when being described differently. The importance of identity resolution has been greatly felt these days in the world of online social networking where personal details can be fabricated or manipulated easily. In this research a new graph-based approach has been used for identity resolution, which tries to resolve an identity based on the similarity of attribute values which are related to different identities in a dataset. Graph analysis techniques such as centrality measurement and community detection have been used in this approach. Moreover, a new identity model has been used for the first time. This method has been tested on SPIRIT policing dataset, which is an anonymized dataset used in SPIRIT project funded by European Union’s Horizon 2020. There are 892 identity records in this dataset and two of them are ‘known’ identities who are using two different names, but they are both belonging to the same person. These two identities were recognized successfully after using the presented method in this paper. This method can assist police forces in their investigation process to find criminals and those who committed a fraud. It can also be useful in other fields such as finance and banking, marketing or customer service.