Discrimination of targets can be improved significantly by analyzing the polarization of scattered electromagnetic waves. In radar imaging, the target image can be enhanced by combining measurements from different polarizations. In this chapter, we propose a joint image formation and fusion approach for multipolarization through-the-wall radar imaging, using compressive sensing (CS). The measurements from different polarization channels are processed jointly using the multiple measurement vector (MMV) model to produce several images of the scene, each corresponding to a polarization channel. Furthermore, the measurement vectors are fused together to form a composite measurement vector, which yields a composite image of the scene. The advantage of fusing the measurement vectors before image formation is that the measurement noise is reduced and the target information is enhanced, which leads to a more informative composite image. The MMV model enforces the same sparsity support for all formed images by reinforcing target information across channels and attenuating noise. Experimental results are presented using simulated and real data. Analysis and comparison of experimental results demonstrate the effectiveness of the proposed through-the-wall radar imaging approach, especially in the presence of high-measurement noise.