This article explores the problem of achieving fixed-time consensus in high-order nonlinear multi-agent systems. A self-regulation hierarchical sliding mode structure with adaptive gains is designed. Different from existing results, the adaptive gains in the proposed self-regulation sliding mode structure can be adjusted automatically by system performance. Meanwhile, a distributed control strategy is proposed using neural networks to approximate nonlinear functions for the discussed problem. The proposed control strategy guarantees that the consensus error converges to the vicinity of zero within a fixed time. Finally, a numerical simulation example is used to verify the superiority of strategy.