Palmprint recognition, noted for its high reliability and privacy has recently aroused wide attention in biometrics. For palmprint recognition, High-quality palmprint images are more effective. However, in real-world scenarios, the quality of palmprint images tends to be low due to limitations in device capabilities and various environmental factors. This study introduced a palmprint recognition method based on image super-resolution (SR), called PPSRNet. The proposed PPSRNet consists of two parallel branches that extract features from SR palmprint images and original images, merging them through score fusion at the final stage. This method employs image SR to enhance image quality and refine the intricate details of palm texture, thereby improving recognition performance. It also utilizes score fusion to address potential negative effects during the image SR process. Experimental results on three popular public palmprint biometric datasets demonstrated the effectiveness of PPSRNet for palmprint recognition.