This paper showcases a framework for predicting hand, foot, and mouth disease (HFMD) risk in southern China by integrating multisource data. The authors used data from Bao’an District in Shenzhen between 2014 and 2023, including HFMD cases, meteorological and air pollution variables, Baidu Index search data, and public health measures. A correlation analysis was conducted to explore the relationships between HFMD incidence and these systematic factors. To capture delayed and nonlinear environmental effects, the study used a Distributed Lag Nonlinear Model. For forecasting, the researchers used a Seasonal Autoregressive Integrated Moving Average (ARIMA) model for 1 week ahead predictions and advanced machine learning methods for 2 to 4 week ahead predictions. They found that all environmental and search behavior factors except sulfur dioxide showed significant nonlinear associations with HFMD incidence. The Seasonal ARIMA model achieved the highest accuracy for 1 week ahead forecasts (R²=0.95, r=0.98). For longer term predictions, models combining web and environmental data had stronger performance and more stable results. The integrated approach matched predicted accurately and observed risk levels which shows the value of combining surveillance, environmental, and digital data for the prediction of epidemics.