Artificial intelligence (AI) is becoming one of the most essential driving forces for 6G communication network and industrial upgrading. The native AI introduces four elements of connection, calculation, model and data resources. The deep integration of heterogeneous resource is an important technical feature of native AI in 6G networks. This paper considers a communication and computing fusion mechanism of AI tasks, which consider the user's performance parameter. The user's performance parameters include processing performance and display performance parameters, which can affect the setting of network parameters such as base station guarantee bandwidth and reliability. We consider both multiple users independently initiate AI tasks and multiple users dependently initiate AI tasks. To provide accurate verification results, we simulate the performance gains of multi-user AI tasks for a specific base station selection. The simulation results show that the proposed solution can maximize the utilization rate of base station resources and maximize the number of users that meet the experience requirement.