Summary: The number of sources must be known for eigen-decomposition subspace direction finding algorithms. However, in the case of small number of snapshots, low signal-to-noise ratio, and there are different signal strengthes between sources, both conventional AIC information criterion and MDL criterion are not able to judge the source number correctly and the performance of eigen-decomposition subspace direction finding algorithms (such as MUSIC method) deteriorates. To solve the problem, a hybrid direction finding approach using priori information of sources is proposed. With the proposed approach, the number of sources can be determined by using the fact that sources are sparse in angle domains. Furthermore, with the priori information that sources are non-circular, the direction finding performance is improved. Finally, computer simulations illustrate correction of the proposed approach.