This paper proposes a multimodal human personal recognition system based on palm print and knuckle print. Palm is the inner surface of the hand that extends from the wrist to the base of the fingers and contains a lot of unique pattern of ridges, valleys, principal lines and wrinkles. On the other hand, Knuckle is the part of a finger at a joint where bone is near the surface. Pattern formation at finger knuckle are unique and hence provide knuckle good discriminative power. Knuckle is rich in texture pattern. In this work edge based local binary pattern (ELBP) is used for image enhancement. Corner features are tracked using LK-tracking constrained by Gaussian response pattern and some geometrical constrains. This system has been tested on some publicly available databases such as CASIA-PALM, Poly U-PALM and Poly U-KNUCKLE. The system has been found to perform well over these databases and fusion has shown significant improvements.