The precise determination of geomechanical parameters has been seen as vital for alleviating the potential problem of the wellbore's instability during drilling operations. In fact, the migration of fine particles can destroy the natural permeability of sediment, a process that can severely impact wellbore integrity. A careful estimation of these components is essential for drilling optimization and the well completion process. Minimum horizontal stress, which one of the three principal stresses, is crucial for assessing both hydraulic fracturing and wellbore stability. This study present a data-driven technique for predicting the minimum horizontal stress in five wells located in the Eocene Agbada Formation of the Niger Delta. Six machine-learning models were applied to the traditional logging data to predict the minimum horizontal stress. The comparison analysis of the applied algorithms shows that the gradient-boosting outperformed the other types. They give the best performance with the highest R2 of 0.92, lowest MAE of 279.66 on the training set, and high R2 of 0.93, and lowest MAE of 273.53 on the holdout set. By doing a blind test on two wells, we were able to confirm the effectiveness of the suggested method. The model was affirmed to be generalized with low generalization errors on the unknown data.