The vehicular edge computing system integrates the computational resources of vehicles, and provides task offloading services for other vehicles and pedestrians. However, while ensuring low response latency for vehicle task offloading, the vehicular network also requires uploading contextual information, such as vehicle speed and location, to road side units and base stations. This introduces significant threats and risks to the privacy of the vehicular network. In this paper, we propose a Task offloading, Data security and Resource allocation offloading algorithm based on the Proximal Policy Optimization method (TDR-PPO). To ensure the security of transmitted data, a suitable encryption algorithm is selected through TDR-PPO for minimizing the average failure probability of all tasks. Experimental results show that the proposed method significantly outperforms benchmark algorithms in terms of average latency and data failure probability, highlighting its effectiveness in balancing performance and security in dynamic vehicle edge computing.