Vortex optical multiplexing communication technology can effectively improve the channel capacity of a communication system in an underwater channel. However, ocean turbulence causes the inter modal crosstalk of the vortex beam, degrading the performance of the communication system. To alleviate the modal crosstalk problem, this study presents a blind equalization algorithm based on a back propagation (BP) neural network. Four channels of vortex light are used for multiplexing transmission, and the random phase screen method is used to simulate ocean turbulence. After the BP blind equalization algorithm is implemented, the improvement in the system bit error rate is simulated and analyzed under conditions of varying ocean turbulence intensities, transmission distances, and vortex light multiplexing modes. The simulation results show that the blind equalization algorithm using the BP neural network can effectively reduce the impact of ocean turbulence on the bit error rate of the system, and the system performance significantly improves when the multiplexing mode interval is 2.