Breast cancer represents the leading cause of fatality among cancers for women and there is still no known way of preventing this pathology. Early detection is the only solution that allows treatment before the cancer spreads to other parts of the body. Diagnosis of breast cancer at the early stage is a very difficult task as the cancerous tumors are embedded in normal breast tissue structures. Aiming to model breast cancer prediction system, we propose a novel machine learning approach based on wavelets. The new model, called wavelet neural network (WNN), extends the existing artificial neural network by considering wavelets as activation function. The texture information in the area of interest provides important diagnostic information about the underlying biological process for the benign or malignant tissue and therefore should be included in the analysis. By exploiting the texture information, a computerized detection algorithm is developed that are not only accurate but also computationally efficient for cancer detection in mammograms. The texture features are fed to the WNN classifier for classification of malignant/benign cancers. An experimental analysis performed on a set of 216 mammograms from screening centres has shown the effectiveness of the proposed method. © Springer International Publishing Switzerland 2014.