In the field of biometrics, palmprint recognition has received great interest and made tremendous progress in the past two decades. In palmprint recognition, the important step is to extract the discriminative features from the image and compare it with templates for identification and verification tasks. In this paper, a new genetic-based 2D Gabor filter with the Convolutional Neural Network is presented. The scale and orientation details captured by Gabor filters are optimized based on central frequency, which is determined based on genetic algorithm fitness function. The proposed technique is implemented on four publicly available palmprint datasets- PolyU, CASIA, IITD, and Tongji. Experimental results confirm that the proposed technique achieves better accuracy when compared to Palmnet.