Recently, Unmanned Aerial Vehicle (UAV)assisted communication and UAV-assisted sensing has attracted great attention due to the excellent performance of high mobility., flexible deployment, and low costs. However, UAV communication loads and radar loads are independent of each other at present, leading to a lower spectral efficiency and a higher hardware cost. With the evolution of integrated sensing and communication (ISAC), the capacities of UAV communication and sensing are in the trend of integration. For ISAC-empowered UAV platforms, the efficient utilization of ISAC signal seems significant to pursue the integration gain. Motivated by this, the intelligent fusion issue of communication signal and sensing signal is mainly discussed at a ISACempowered UAV platform in this paper. We firstly propose an ISAC signal compression framework based on Generative Adversarial Network (GAN) to reduce the information entropy of the input signal. Then two ISAC signal fusion frameworks are designed based on deep semantic matching and multi-layer semantic matching, respectively. The proposed intelligent fusion frameworks can efficiently fuse the communication signal and sensing signal for the further information sharing between UAVswarms.