The precise classification of jamming signals holds paramount significance in the effective implementation of anti-jamming strategies within communication systems subject to intricate environmental variables. In light of this imperative, we propose an innovative fusion algorithm based on conditional generative adversarial network (CGAN) and convolutional neural network (CNN) to solve the problem of difficulty in applying deep learning (DL) algorithms due to the instantaneous nature of jamming signals in practical communication systems. Compared with previous methods, our algorithm achieved an 8% improvement in accuracy even when working with a limited dataset. Unlike previous research, we have simulated real-world satellite communication scenarios using a hardware platform and validated our algorithm using the resulting time-domain waveform data. The experimental results indicate that our algorithm still performs extremely well, which demonstrates significant potential for practical application in real-world communication scenarios.