The vectorial total variation model and algorithm are studied for multi-channel SAR image denoising. After introducing the vectorial total variation model, an accelerative fix-point iterative algorithm was proposed and its convergence was proved. By improving the filter coefficient of the fix-point iterative process, an adaptive vectorial total variation model was developed for multi-channel SAR image denoising, whose iterative algorithm and convergence theorem were present. The performance of denoising and resolution preservation of our models was tested by multi-polarimetric, multi-temporal RADARSAT-2 images, in addition to the validation of the convergence and the convergent speed of the proposed algorithms.