In this article, a tensor-based angle estimation algorithm is proposed for bistatic multiple-input multiple-output (MIMO) radar systems with a multislot gain-phase error (MSGPE). First, the received signals are constructed into a complex-valued third-order parallel factor model that contains information, such as transmit MSGPE, receive MSGPE, directions of departure (DODs), and directions of arrival (DOAs). Then, a two-stage tensor-based angle algorithm is proposed for antenna arrays with the MSGPE. In the first stage, the MSGPE information part and angle information part are separated by the least-squares Khatri–Rao factorization. In the second stage, the DODs and DOAs are obtained jointly based on spatial smoothing and real-valued processing techniques under eliminating the influence of MSGPE. Compared with the existing competitive algorithms, the proposed algorithm achieves an improved performance and is applicable for both the coherent and noncoherent targets. Simulation results show the superiority of the proposed tensor-based angle estimation algorithm.