This paper researches on millimeter-wave (mm-wave) channel estimation based on interference cancellation (IC) and DOA (Direction of Angles) estimation in hybrid massive MIMO systems. According to the ray-tracing model, the mm-wave channel estimation problem can be solved by converting into the estimation of parameters including the angles of departure (AOD), the angles of arrival (AOA) and the path gains. In order to separate different paths in the mm-wave channel, interference cancellation is carried out by utilizing a designed set of spread codes. Then the DOA estimation and the least-squares method are used to achieve the path directions and gains, respectively. Compared with estimating all the paths directions simultaneously, simulation results show that the combination of IC and MUSIC can get lower root mean square error (RMSE). Meanwhile, the algorithm can get high resolution in directional spectrum because the proposed method can suppress the additive noise in the mm-wave channel effectively.