The COVID-19 pandemic has introduced unprecedented challenges to global mental health, with anxiety disorders emerging as a prevalent concern. This research study aims to predict the severity of anxiety using advanced machine learning techniques by analyzing the dataset collected from Statistics Canada that include behavioral and demographic variables, indigenous identity, labor market variables, mental health variables, and sociodemographic characteristics. This study attempts to expand on our knowledge of the pandemic’s effects on mental health by identifying important predictors of anxiety severity. The outcomes of this research are anticipated to help advance proactive approaches, providing tailored psychological assistance and focused treatments to efficiently reduce anxiety during and after the pandemic. This work not only bridges the gaps left by the COVID-19 pandemic but also uses machine learning to improve the predictive power of mental health assessments, which could lead to advancement in more flexible mental health service frameworks.