Although offloading tasks to remote cloud servers can reduce the computation time of tasks, the long distance between user devices and cloud servers also raises the data transfer time and introduces unpredictable network latency. The introduction of edge servers further improves the execution efficiency of user tasks and relieves the pressure on cloud servers. However, the explosive growth of tasks also gradually puts pressure on the edge servers with limited computing resources, which not only affects the execution efficiency of each task, but also causes unbalanced load on all edge servers, thus affecting the efficiency and stability of the edge servers. In addition, with the increasing awareness of user privacy protection, it has become a challenge to maximize the privacy of user data while improving the execution efficiency of task and ensuring the load balance of edge servers. To address this challenge, we design a task offloading method in edge cloud environment, namely NTO. First, the task offloading problem in an edge cloud environment is modeled as a multi-objective optimization model with task execution time, load balancing of edge servers, and privacy conflict of user data as three key optimization objectives. Then, the model is optimized by the Non-dominated Sorting Differential Evolution (NSDE) algorithm. Finally, We conducted experiments to demonstrate the effectiveness and superiority of NTO.