This paper proposed a novel algorithm for combination of two adaptive filters which stem from different families. The algorithm extended the combination coefficient from a scale in traditional convex and affine combination methods to a vector. These improvements would significantly enhance the performances than any other existing methods under various interference environments, because the new method took more parameters in vector into account, which may accelerate convergence speed and enhance the Signal to Interference Noise Ratio (SINR) at output. Furthermore, the paper exploited the Multistage Nested Wiener Filter (MNWF) and LMS for combination according to their obvious complementary characters for anti-interferences and effective implementation of their low computational complexity. Finally, we simulated the proposed algorithm in navigation multiple antenna array application, which have better combination performance than other approaches.