In practical applications, the array received data often contains target signal components, using the covariance matrix of the sample including the partial target signal instead of the covariance matrix of the interference noise will cause significant errors, especially when the number of array received snapshots is low. At the same time, the target signal is often weak, and direct angle measurement will result in certain angle measurement errors, leading to errors in the target signal guided vector. These two non-ideal application scenarios will seriously deteriorate the algorithm’s anti-interference performance. This paper proposes an anti-mainlobe interference algorithm based on robust oblique projection preprocessing. Firstly, the interference noise covariance matrix is reconstructed based on the received data including the target signal, to minimize the impact of the target signal on the algorithm performance furthest. Secondly, based on the reconstructed interference noise covariance matrix and the high-precision estimated target steering vector, a robust oblique projection preprocessing matrix is jointly designed to effectively suppress the mainlobe interference. Finally, the data after oblique projection preprocessing is processed by robust adaptive beam nulling to suppress the sidelobe interference. Simulation results show that the proposed algorithm can effectively suppress the mainlobe interference and sidelobe interference while eliminating the mainlobe distortion, and has good robustness to the non-ideal scenarios of the received data, which includes the target signal and exists angle measurement errors.