This study investigated the influence of thermal environment on the drinking behavior of beef cattle using electronic water troughs. Forty-six weaned male Caracu calves (mean±SD: 247±230 days old, 226±35 kg body weight) were evaluated over a 68-day feed efficiency test during the dry season in São Paulo State, Brazil. Housed in a collective pen, the calves had access to 12 electronic feeders and four water troughs. The system recorded the number of visits to water troughs, duration of drinking events (in seconds), and water intake (kg). The drinking behaviors were summarized by day to create one observation per day per animal. The Temperature Humidity Index (THI), adjusted for wind speed and solar radiation, was calculated using data from a local weather station. Daily mean THI and 3-day moving averages of daily mean and maximum THI values were calculated. Linear regression models were used to determine which THI measure best-predicted drinking behavior. Random effects of day and individual animals were incorporated to refine the model fit. All analyses were performed using R, with model selection based on the lowest Akaike information criterion (AIC) and highest R² value. The results indicate that the daily mean THI was the best predictor of drinking behaviors, though it had no significant effect on the number of visits (p=0.55) or event duration (p=0.90). The number of visits was better explained by animals (p<0.05, R²=0.31), while the duration of events was better explained by day (p<0.001, R²=0.43). However, daily mean THI influenced water intake (p<0.001); for each unit increase in THI, calves were 7% more likely to increase their water intake. Overall, while water intake was sensitive to changes in daily mean THI, the number of visits and event duration were better explained by day and by animal factors.