In the wavelet shrinkage method, determining the right threshold is important, especially when estimating the mean matrix parameter for matrix-variate aporically distributions is of interest. This paper introduces a new soft thresholding wavelet shrinkage estimator based on Stein's unbiased risk estimate (SURE) for matrix-variate spherically distributions. Our goal is to find a class of the soft thresholding wavelet shrinkage estimator using a balanced loss function and evaluate its performance through simulation study.