The performance of the underwater wireless optical communication (UWOC) system is severely affected by the presence of air bubbles in the water. Therefore, an experimental setup is designed to investigate the effect of different sized air bubbles on the laser beam propagating through the UWOC system. A wide range of air bubbles has been generated in the water tank by varying the diameter of the holes in a copper tube and by varying the air flow rates. When an optical beam propagates through the underwater environment in the presence of air bubbles, the received optical beam undergoes severe intensity fluctuations. The distribution of these intensity fluctuations conveys very useful information to predict the behavior of the underwater channel. The distributions of these intensity fluctuations have been modeled using a Gaussian mixture model (GMM), which is the sum of Gaussian functions. The parameters of the proposed GMM model are estimated by expectation maximization algorithm to obtain maximum likelihood estimation. The goodness of fit is also performed by considering a confidence interval of (95%) to estimate the feasibility of the proposed GMM model, which provides excellent results. The proposed GMM model excellently fits the experimental data for all the considered cases. In addition, based on this proposed model and experimental observations, the performance of the UWOC system is evaluated in terms of signal-to-noise ratio, outage probability, bit error rate, maximum Q-factor, etc. The results show an exact match between the experimental and proposed theoretical results that are modeled using GMM model and hence signifies the precision of the proposed model.