In this paper, an imaging algorithm for the airborne radar system maneuvering along an arbitrary trajectory is proposed. The algorithm aims at wide-angle imaging with incomplete measurements from the nonlinear trajectory. The proposed composite joint sub-aperture imaging algorithm provides high reconstruction quality and supports efficient data collection policy. The image can be reconstructed by combining image patches corresponding to non-overlapping sub-apertures. The image patch is obtained by compressive sensing with joint sparse representation of the scene. Numerical results have proved that the proposed algorithm is highly effective and capable of image reconstruction without much loss in quality, especially on objects signature and contour.