The adult population has experienced low back pain at some time in their lives, especially in women between 40 and 80 years of age. Given this, the present study aims to compare Machine Learning algorithms for the identification of pain caused by lumbar disc herniation, for which several models were evaluated using key metrics such as Sensitivity (Recall), Classification Accuracy (CA), F1 Score, Area Under the Curve (AUC) and Matthews Correlation Coefficient (MCC). The results indicate that the Neuronal Network model stood out for its sensitivity with a recall of 1.000, which means an excellent capacity to identify positive cases of low back pain, as well as an AUC of 0.500 and MCC of 0.000; Gradient Boosting showed a high sensitivity Recall of 0.999, competitive F1 Score of 0.799, AUC of 0.497 and MCC of 0.002; Random Forest and k-Nearest Neighbors (kNN) obtained moderate results with metrics such as F1 Score around 0. 721–0.722 and Recalls of about 0.786–0.789, AUCs close to 0.5 and low MCCs (0.002 and 0.001 respectively); AdaBoost and the Logistic Regression model showed inferior performance compared to the previous ones, with AUC values around 0.502–0.570 and MCCs close to zero, indicating a limited ability to identify cases of low back pain correctly. In conclusion, we recognize the need to optimize and further explore the algorithms to improve the accuracy in identifying low back pain caused by disc herniation.