Because MOSS (Microsoft Office SharePoint Server) doesn’t support the association between two heterogeneous entities, we propose a clustering algorithm called MHEA. This algorithm can mine the associations among heterogeneous entities and output entity association matrices. It is based on Extended Name Vectors which is the result of ENV clustering algorithm. ENV is based on k-means, but it selects initial centroid of each clustering according to the identifiers of entities, and the value of k is dynamically changeable in the process of ENV, consequently decrease the iterative times. The accuracy and effectiveness of our proposal have been demonstrated through the performance study using both real and synthetic data sets.