is paper proposes feature based image segmentation by combining M-band wavelet transform with fuzzy rule. M-band wavelet transform decomposes an image in to MxM channels. A different combination of these band pass sections produces various scales and orientations in frequency plane, which yields a sixteen sub-band images. The texture features are obtained by subjecting band pass section to the non-linear transformation and by computing the measure of energy in a window around each pixel for the filtered images. Fuzzy rule is a key tool for expressing our piece of knowledge. Fuzzy rule is then constructed for the texture features. Performance of the proposed method is analysed using deviation error, and found that the algorithm produces good segmentation results by generating Fuzzy rules for M-band wavelet derived features