For different social network analysis measures and metrics, the output always obtains a rank ordering of the nodes in terms of these measures. Information including in this kind of results often has highly practical value. However, different measures and metrics always indicate different key nodes. This paper aims to propose a new measure of influence of nodes in social network. Entropy-weighted TOPSIS method is applied to the evaluation of influence of nodes in social network. The weights of centrality metrics are gained with the method of entropy. Then,the TOPSIS method is adopted to research the influence evaluation of nodes in social network. Finally, the results are illustrated by using a network from web of SINA.