This paper presents a combination of palm based contour let and ICA multimode fusion recognition algorithm. The algorithm using mainly the palm and knuckle print section. Firstly it transforms the source images of plamprint and knuckle print by contour let wavelet．After the image transformed by contour let，the low － frequency can reflect the texture profile，high frequency directional sub band can reflect the contour，texture and so on. So the extracted can conduct dimensionality reduction for data by ICA algorithm. Then it implements the normalized and simple weighted fusion．Finally，the classification than obtained the recognition rate. Experimental results show that the fusion system recognition performance is superior to the performance of a single plam print recognition or knuckle print，thus proving the validity of the algorithm.