A new data association algorithm is proposed in this paper for heterogeneous sensors information fusion system, which consists of Radar and high-dynamic-range IR, based on joint usage of the nearest-neighbor (NN) and the topology similarity. The proposed algorithm can avoid the complicated system biases compensation based on topology information. Because the NN method is sensitive to the system bias while the existing topology-based algorithms need a large amount of computation, the proposed algorithm shows a lot of advantages, like high estimation accuracy and robustness as well as the improvement of success association rate. Finally, some numerical experiments are provided to demonstrate the effectiveness of the proposed method. And the estimation accuracy can be as high as 90%.