Breast density is a widely adopted measure for early breast cancer diagnose. In this paper, an automated breast density estimation method was proposed. Mammograms were analyzed using wavelet transform to extract tissue-like contents. A tissue image was then divided into fixed size sub-regions. The sub-regions were classified as high and low density categories using their distribution features. In this paper, groups of histogram moments were extracted as features of sub-regions, and served as inputs of the support vector machine (SVM) for classification. The breast density of the whole mammogram was then evaluated by calculating the ratio of number of high density sub-regions to that of the whole set. Experimental results show the excellent performance of the proposed method.