The present work aims at comparing the role and efficiency of two artificial neural networks, which are logistic regression and decision tree models for the detection of TB positivity depending on demographic factors and symptoms. It has also provided the patient details such as age, gender, and different symptoms of TB. Both and were split into training and testing sets with the assessment of the results’ accuracy based on the comparison of the results to the data, confusion matrices and ROC curves. Logistic regression model yielded an accuracy of 90% while decision tree yielded an accuracy of 85% as shown by the results above. The ROC curves presented herein showed a good predictive ability of both models where AUC values of 0.85 for the Naive Bayes and 0. Based on the results obtained from the experiments, we have identified that the accuracy achieved for this model is 92 for the decision tree. Even though, in this case the logistic regression’s accuracy is slightly higher than the decision tree, both types of models offer a good prognosis for TB diagnostics. The findings of this comparative analysis support the consideration of applying machine learning for the improvement of the accuracy in diagnosing TB and presents future directions for this line of research.