To reduce the influences of noise and time delay on positioning accuracy in indoor visible light positioning, a new indoor light-emitting diode (LED) visible light positioning algorithm is proposed. A honeycomb topology structure is introduced to locate the indoor visible light, and its beacon node is selected adaptively. The influence of the noise on the received optical power is studied, and a mean model is used to correct the received optical power. However, the time delay of conventional time division multiplexing system is too long, and thus the clock synchronization error occurs easily. In order to meet the high precision real-time positioning, an improved time division multiplexing network system is used. A network source center node is selected through the cell division. The partial time slot is reused through time slot competing so that the contradiction between the frame length and the number of LED nodes is solved. The simulation results show that the proposed algorithm can effectively reduce the influences of target moving and clock synchronization error on positioning accuracy, and improve the real-time performance and the stability of the indoor positioning system.