The virtual reality engine in the Metaverse has performance bottlenecks in terms of rendering quality, rendering efficiency, and data transmission. This paper uses viewpoint-based intelligent slicing transmission technology to dynamically transmit required data according to the user's perspective, reducing unnecessary data transmission; the paper applies ATW technology to increase the frame rate by inserting the previous distorted frames, reducing the hardware requirements; this paper adopts asynchronous space warp technology to solve the frame jitter problem by filling the missing frames, generating new frames and preventing the rendering of duplicate frames; finally, through parallel processing technology, the game thread and the rendering thread are executed in parallel to save rendering time, and the fade-in and fade-out detail level algorithm is used for the first layer of clipping, and then the slow culling algorithm is used for the second layer of clipping. The research results of this paper show that through the application of digital intelligent network technology, the performance of the virtual reality engine in the Metaverse has been significantly optimized. Experimental data show that the optimized frame rate increased from 121 FPS to 150 FPS to 152 FPS to 205 FPS, the rendering time decreased from 669 ms to 899 ms to 420 to 640 ms, and the packet loss rate decreased from a maximum of 4.06% to 2.49%. The optimization framework and solution proposed in this paper provide a practical path for the development of virtual reality technology in the metaverse.