Abstract: This paper aims at cognitive MIMO system which is composed of on primary link and multiple secondary user interference channels, first proposes an iterative cognitive interference alignment optimization algorithm based on signal subspace and proves the convergence of the proposed algorithm with the monotonic boundary theory. In order to further improve the performance of system’s sum rate, an enhanced cognitive interference alignment algorithm based on joint signal subspace and power allocation is proposed. Due to the adaptive power allocation between multiple data streams of each secondary user, the enhanced cognitive interference alignment algorithm can further relieve interference signals on secondary users’ desired signal subspace. From the simulations, we confirm that our proposed algorithms provide a substantial performance gain over the conventional cognitive interference alignment algorithms.