Mental health disorders are now a most important public health challenges among individuals with a history of drug abuse. Nowadays, machine learning (ML) has apparent as a powerful apparatus in the area of mental health care, necessitating precise predictive models of mental health among drug-abused individuals, especially after the post-pandemic. Here we are taking data from people from Dhaka, Bangladesh, between ages group of 15–48 years. The dataset contains a comprehensive set of parameters including Age, Gender, Education, Living Situation, motive for Drug use, emotional issues, suicidal tendency, family financial condition, effect of substances addicted persons in family, Cohabitation with Drug Users, Smoking Habits, Drug Use History, Influence of Friends etc. The main contribution is that we do cross-performance analysis of Various machine learning algorithms, including Logistic Regression (LR), K-Nearest Neighbors (KNN), Decision Tree (DT), Naïve Bayes (NB), and Support Vector Machine (SVM) algorithms into the same dataset to predict the mental health conditions of drug-adopted individuals. Amid all support vector machine (SVM) algorithm gives the best performing accuracy of 97.7% and for others also has an accuracy greater than 80%.