Abstract: Hydrogen embrittlement (HE) poses a serious threat to the structural integrity and service reliability of low-alloy steels. It is one of the key factors limiting their long-term performance in critical engineering applications. The diffusion, trapping, and accumulation of hydrogen atoms within materials lead to microstructural embrittlement and crack initiation, significantly reducing ductility and fracture toughness. HE behavior is typically governed by the coupled effects of microstructure and environmental parameters, such as temperature, pressure, and hydrogen concentration, and therefore exhibits high nonlinearity and complexity. However, experimental studies are often costly, time-consuming, and difficult to reproduce, resulting in limited datasets. Issues such as data scarcity and uneven feature distribution are common, making it difficult for existing machine learning models to achieve accurate predictions under limited-sample conditions. In recent years, data augmentation techniques have been increasingly adopted in materials science to address data scarcity under small-sample conditions. Data augmentation methods expand datasets by statistically perturbing and synthesizing original samples while preserving the data distribution. These methods have shown promising results in alloy design, fatigue life prediction, and corrosion modeling. To address HE in low-alloy steels, this study proposes a quantile Gaussian data augmentation with multi-model learning (QGDAM) method for small-sample HE behavior, aiming to achieve robust learning and high-precision prediction under limited-data conditions. The method consists of two modules: data augmentation and regression prediction. In the data augmentation phase, a quantile transformation module is introduced to mitigate skewed feature distributions. Three data augmentation strategies based on Gaussian mixture models (GMM) are designed to generate augmented samples that closely match the distribution of the original data. In the prediction phase, a multi-model ensemble regression framework is established, incorporating random forests, gradient boosting, light gradient boosting (LightGBM), and k-nearest neighbors. This study extracted 90 valid samples from publicly available HE experimental data to construct a low-alloy steel HE behavior database containing 17 key features. These features encompass material strength parameters, environmental conditions, and elemental composition. The results demonstrate that the QGDAM method significantly outperforms traditional machine learning approaches in both sample quality and prediction accuracy. Compared with existing methods, this study demonstrates higher sample quality across four sample augmentation ratios. Compared with the baseline model, QGDAM achieves a significant reduction in mean squared error (MSE) and an average increase of 0.18 in the coefficient of determination ( R^2 ). Additionally, the method significantly improves prediction accuracy on two external validation sets, indicating strong generalization capability and robustness. Furthermore, this paper compares the feature-response relationships of models before and after augmentation using partial dependence plots (PDP) and Shapley additive explanations (SHAP). The results show that the augmented model more accurately captures the influence patterns of changes in feature variables on HE sensitivity. In contrast, the dependency curves derived from the original dataset exhibit scattered distributions and weaker regularity, indicating that QGDAM significantly enhances the model’s ability to fit real physical mechanisms at the feature-learning level. Comprehensive results demonstrate that the proposed QGDAM method effectively improves the accuracy and interpretability of HE behavior prediction under small-sample conditions. This provides a generalizable data-driven approach for intelligent modeling of the service performance of complex materials.