During the development of natural gas reservoirs, predicting the bubble size distribution and transport velocity in the study of gas-water two-phase flow through porous media is extremely important in natural gas recovery. Considering the different forces that act on small and elongated bubbles, a double-bubble phase population balance model was introduced to establish an improved bubble flow model. Accordingly, a two-dimensional visualized model was designed to analyze the bubble diameter distribution and average velocity of bubbles. Dominant bubble coalescence, breakup, and coexistence flows were simulated to evaluate the uniqueness of the diameter distribution of bubbles, the change in the median bubble size at different dimensionless distances, and the average velocity distribution. The peak distribution of the bubble diameter was different in the three flow conditions. Owing to the difference in the bubble breakup and coalescence rates, the peak values of bubbles varied with the increase in the dimensionless distance. Higher bubble breakup or coalescence rates were accompanied by a rapid change in the median bubble size. Furthermore, the average velocity of the transition from small bubbles to elongated bubbles decreased significantly. Thus, the proposed pore-scale gas-water two-phase model can effectively predict the occurrence and transport processes in natural gas reservoirs.