Forecasting mosquito breeding sites is essential for guiding vector control strategies in endemic malaria and dengue regions, including Sudan. Time series models provide a valuable approach for capturing seasonal patterns and predicting fluctuations in breeding activity. Data on mosquito breeding sites in Khartoum State from 2001 to 2022 were analyzed using the Box-Jenkins methodology. After ensuring stationarity through logarithmic transformation, first differencing, and seasonal differencing, candidate SARIMA models were identified based on autocorrelation and partial autocorrelation plots. Competing models were evaluated using fit statistics, residual diagnostics, and the principle of parsimony. Among the tested specifications, SARIMA (1,1,1)(0,1,1)₁₂ emerged as the most appropriate model, achieving a stationary