Integrated sensing and communication (ISAC), as a key technology for 6G communication, alleviates frequency spectrum pressure by enabling communication and sensing to share the same hardware platform and frequency spectrum resources. However, high-frequency signals exhibit weak penetration capabilities. Particularly in non-line-of-sight (NLoS) scenarios where obstacles exist between the transmitter and receiver, the performance of communication and sensing sharply decline. To solve this problem, reconfigurable intelligent surface (RIS) has been widely studied for its ability to improve signal attenuation and communication environment. This paper constructs a new channel model using unmanned aerial vehicles (UAV) equipped with RIS to solve the problem of NLoS communication faced by ISAC systems. The beamforming weights of base station (BS) and phase shift of RIS are jointly optimized by deep neural network (DNN), so as to maximize the user’s communication sum rate and sensing effect on sensing target. Simulation results demonstrate the algorithm proposed in this paper, namely ISAC-Aerial RIS-DNN (IARD), is effective for NLoS communication and achieve higher sum rate.