The establishment of steel fiber-reinforced ultra-high-performance concrete (SFUHPC) signifies a significant breakthrough in concrete technology, showcasing superior performance compared to conventional fiber-reinforced cementitious composites. SFUHPCs not only boast higher tensile strength but also exhibit greater tensile ductility under both static and dynamic loads, attributed to the unique strain-hardening behavior accumulated in multiple microcracks. Consequently, this material holds promise for structural applications under quasi-static and dynamic loading conditions. While several empirical models have been developed to forecast the dynamic increase factor (DIF) of SFUHPC tensile strength, their limitations arise from being based on a fixed set of individual test data, potentially lacking the necessary breadth to ensure accuracy. This study introduces a novel approach by employing a machine learning model to estimate the DIF of SFUHPC tensile strength. Drawing from a collection of 140 experimental test results, an artificial neural network (ANN) model has been developed, accounting six input parameters characterized by intricate relationships. The prediction results underscore the excellent performances of the developed ANN model in forecasting the DIF of SFUHPC tensile strength across both training and testing sets. Through a sensitivity analysis, it was revealed that the strain-rate emerged as the greatest contribution parameter, significantly impacting the DIF of SFUHPC tensile strength, while fiber length exhibited the least importance among the variables.