The development of a quantitative diagnostic method for liver fibrosis using an ultrasound B-mode image is highly required. In our previous study, a multi-Rayleigh model was proposed to express a probability distribution of echo envelope amplitude from a fibrotic liver. Using the multi-Rayleigh model, a structure of fibrotic tissue can be quantitatively estimated. In this study, a method of estimating the number of tissue components was proposed to improve the accuracy of estimating a fibrotic tissue structure. Using threshold processing for a squared Mahalanobis distance of moments, which is a statistical property of echo envelope amplitude, the number of tissue components could be quantitatively estimated. The results of evaluation of clinical ultrasound B-mode images using the multi-Rayleigh model with estimation of the number of tissue components well reflected the tissue structural changes caused by liver fibrosis. It was concluded that our proposed method of estimating the number of tissue components improves the accuracy of liver fibrosis evaluation based on the multi-Rayleigh model.