Membrane dispersion has extensive application prospects in chemical and petrochemical industries, water treatment, and various other fields. The membrane flux serves as a pivotal parameter for assessing the production efficiency, optimization, and scalability. Nevertheless, establishing a mathematical model that accurately describes the relationship between the membrane flux and its associated factors is a challenge. Herein, the flux variation of multichannel ceramic membranes during bubble preparation was investigated by experimental analysis and mathematical modeling. The membrane flux is significantly influenced by factors such as the membrane pore size and membrane channel number. Additionally, the system pressure, gas−liquid flow ratio, liquid viscosity, and surface tension exert substantial effects on membrane flux fluctuations. A prediction model of the membrane flux is established based on the empirical model for the first time, and its relative error is within 10%. The model can predict the optimized membrane area and membrane pore size across different operating conditions and liquid properties under a given membrane flux. This work will contribute to advancing our understanding and development of the bubble preparation process through membrane dispersion.