Abstract: When the interpolation technology is applied to beamforming, data for a virtual array are required to be obtained using interpolated transformation of real data. However, the interpolation mapping errors will affect the performance of virtual array beamforming, especially at wide angles. In this paper, we present an improved algorithm by optimizing the array weights to constrain the mapping errors. The key feature of this algorithm is keeping the orthogonality between the mapping errors and the weights, which can be achieved by loading the covariance matrix of the mapping errors on that of the signals. By doing this, the beamforming performance of the virtual array can be improved, even for relatively large mapping errors. Numerical simulations confirm the validity of proposed algorithm for wide angle beamforming, which shows stable and deep nulls in the direction of interference, and thus, increases the signal to interference and noise ratio (SINR) of the virtual array beamforming. In comparison with the existing methods, the proposed virtual array beamforming method can compensate the real array errors, and be feasible over wide sectors as well, with lower computational complexity.