CO₂ flooding is greatly affected by areal heterogeneity in low permeability reservoirs，causing uneven pressure distribution，early gas breakthrough and low recovery factor. This study explored the applicability of diffluence management to improve CO₂ flooding under laboratory condition using physical experiment simulation. Based on the permeability ratio of the actual reservoir block，a radius model with areal heterogeneity was made to simulate CO₂ flooding as five spots pattern and relative experimental methods were designed. Three production patterns，including constant pressure，preproduction in low permeability well and high producing pressure differential in low permeability well，were introduced to evaluate the gas flooding effects. Results showed that under constant pressure production pattern，there was a significant gas breakthrough in only one production well located in high permeability area，and a huge production difference over the other four production wells，and only 23.11% oil was recovered. The latter two methods produced more oil and delayed gas breakthrough time，achieving oil recovery 42.50% and 46.42% ，respectively. By high producing pressure differential in low permeability well，the injected gas was evenly distributed across the radius model，leading to a lager gas sweeping volume，a higher gas utility and more oil production. Because of low operational cost，diffluence management method is expected to be an effective way to improve CO₂ flooding efficiency in low permeability reservoirs with areal heterogeneity