Differential pressure of gas-liquid mixture contains abundant information about fluid flow and its nature. In this paper, a differential pressure across a Venturi tube is measured for a gas-liquid mixture. Trend component and fluctuation components of the differential pressure are extracted using Extreme-Point Symmetric Mode Decomposition. The analysis shows that the mean of the trend component is related to liquid flowrate, and the amplitude of the fluctuation components is correlated to voidage and flow pattern. Hence, a fluctuation coefficient based approach to access voidage is proposed. This is based on the trend and fluctuation components. Since there exists an influence of gas/liquid flowrate, fluid density and voidage on fluctuation coefficient, their qualitative relationships are analyzed to find the appropriate variables to modify the fluctuation coefficient. Experimental data are used to determine the appropriate specific modification parameters. The modified fluctuation coefficient is found to be flow pattern dependent, and hence fuzzy pattern recognition is adopted to identify flow patterns combining statistics from differential pressure. Finally, a flow pattern-based correlation is proposed to estimate the voidage. Verifications through confrontation with experimental results show that the proposed correlation is effective in estimating voidage of mixtures. (C) 2018 Elsevier Ltd. All rights reserved.