Geogrids are commonly used in geotechnical engineering as soil reinforcement in various applications, such as retaining walls and slopes. The pullout resistance factor (F*) is an essential parameter that governs the performance of the geogrid-reinforced soil system. This study demonstrates the application of a machine learning (ML) algorithm to predict the F* of geogrids. The decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost), and multi-layer perceptron (MLP) models were trained on 550 data points collected from a detailed literature survey. The input parameters included vertical effective stress ( \(\sigma_{{\text{v}}}^{\prime }\) ), relative compaction (RC), the embedded length of reinforcement (Le), fines content (FC), the ultimate tensile strength of geogrid (Tult), pullout strain rate (SR), and the ratio of the spacing between longitudinal and transverse geogrid members to the average particle size of the soil (Sl/D50, St/D50). The performance of these algorithms is evaluated by various statistical measures, including root mean squared error (RMSE), mean squared error (MAE), and coefficient of determination (R2). The results show that the ensemble-based ML algorithms accurately predict F* with a high degree of accuracy (RMSE = 0.19 and R2 = 0.84). Furthermore, sensitivity analysis was carried out to evaluate the most influencing soil parameters affecting the prediction of the pullout resistance factor. The results show that effective vertical stress has the greatest influence, followed by the relative compaction, and embedment length, whereas Sl/D50 has the least importance. In conclusion, the proposed ML algorithm can be an effective tool for predicting the F* of geogrids in sands.