In order to improve the signal-to-noise ratio and mitigate the damage of noise to the communication quality, an effective interference suppression algorithm, which is based on the improved density-based spatial clustering of applications with noise algorithms, is proposed for visible light communication systems using the complementary metal–oxide–semiconductor image sensor as the receiver. Experimental results show that the algorithm can learn the region where the payload data is located, filter out the noise data, and greatly decrease the interference. The effect of the algorithm is also studied through bit error ratio performance.