Matrix completion is widely used in information science fields such as machine learning and image processing. The alternating direction method of mul- tipliers (ADMM), due to its ability to utilize the separable structure of the objective function, has become an extremely popular approach for solving this problem. But its subproblems can be computationally demanding. In order to improve computa- tional efficiency, for large scale matrix completion problems, this paper proposes an improved ADMM by using convex combination technique. Under certain assump- tions, the global convergence of the new algorithm is proved. Finally, we demonstrate the performance of the proposed algorithms via numerical experiments.