The COVID-19 pandemic has presented an unprecedented challenge to the education system, necessitating data-driven strategies to mitigate its impact on students and staff. This research paper introduces a novel machine-learning approach for forecasting COVID-19 hotspots in K-12 schools across Florida. Our study leverages comprehensive datasets encompassing epidemiological, environmental, demographic, and school-specific factors. This research paper showcases a machine-learning approach for forecasting COVID-19 hotspots in Florida’s K-12 schools. By harnessing the power of data and predictive analytics, our approach in this paper empowers education stakeholders to proactively manage and mitigate the pandemic’s impact. Our preliminary results are promising. The four machine learning models (Logistic Regression, Support Vector Machine, Random Forest, XGBOOST) have demonstrated their ability to identify potential hotspots and provide valuable lead time for proactive interventions. This research represents a critical step in enhancing the safety of Florida’s public schools during the ongoing pandemic. This research contributes to machine learning and public health and is a vital tool in the ongoing battle against COVID-19 in educational settings.