In indoor Visible Light Communications (VLC) network with multiple antennas and multiple users, in order to improve the time complexity of optimal antenna selection algorithm under the condition of increasing number of transmitting antennas and users, Naive Bayes (NB) method is applied to the downlink Light Emitting Diodes (LED) selection problem of indoor multi-user VLC network. Firstly, the LED selection task was modelled as a multi-classification problem, and the training sample set was generated by using the users’m known channel state information. Then, the corresponding class labels were generated by maximizing the sum-rate of multi-user VLC network. Secondly, using the generated training sample set, the classifier model is obtained by NB method. Finally, the trained classifier model is applied to the LED selection of new user. Simulation results show that, compared with the optimal LED selection algorithm of multi-user VLC network, the proposed LED selection scheme aided by NB can effectively reduce the time complexity and achieve a good balance between algorithm complexity and system sum-rate.