Texture is an important image feature. A novel texture feature extraction technique is proposed based on coefficient co-occurrence matrix of discrete wavelet frame transformed image, which captures the information about relationship between each high frequency subband and low frequency subband of the decomposed image at the corresponding level. It is not independent to extract the information of each subband coefficient. Considering that the Support Vector Machine (SVM) has advantages of resolving the small-sample statistics and generalizing ability, the classification performance is analyzed by using the SVM classifier. The experimental results demonstrate the effectiveness of our proposed texture feature in achieving the improved classification performance.