<p>Fluid identification of reservoirs is remarkably significant for the evaluation and development of tight sandstone oil gas reservoirs. Tight sandstone reservoirs are characterized by low porosity, low permeability, and strong heterogeneity, leading to complex oil–water relationships. To address the low accuracy of conventional log interpretation methods in identifying fluid types, the author proposes an intelligent fluid identification method based on the Chaos Sparrow Search Optimization Algorithm-Double Cost Sensitive Random Forest (CSSOA-DSRF) model. The Double Cost Sensitive Random Forest (DSRF) incorporates cost-sensitive learning into both the feature selection stage and the ensemble voting stage of the random forest algorithm. By assigning weight coefficients to different fluid types, DSRF enhances the model’s attention to minority class samples, making feature selection more targeted. Subsequently, a set of decision trees more sensitive to minority class data is selected, effectively addressing the issue of class imbalance. Finally, the Chaos Sparrow Search Optimization Algorithm (CSSOA) integrates an improved Tent chaotic map and a Gaussian mutation mechanism into the framework of the Sparrow Search Algorithm (SSA). This enhances population diversity and global search capability while reducing the risk of premature convergence. Based on principal component analysis, nine logging curves, namely acoustic, spontaneous potential, density, caliper, neutron, natural gamma ray, and resistivity logs (AT20, AT60, and AT90), were selected as characteristic parameters. The results indicate that the proposed method achieves a prediction accuracy of 92.13%. Compared with DSRF, Random Forest (RF), K-Nearest Neighbors (KNN), and Support Vector Machine (SVM), the new method demonstrates higher accuracy, low error, along with robust stability and reliability, offering a feasible solution for fluid identification in tight sandstone reservoirs.</p>