The optimization of resources focused on hospital stays plays a critical role in healthcare efficiency issues. To be exempt from proper hospital resource planning puts a nation's health factor at risk. The Additive Winters forecast model has presented favorable and accurate results in the medical field. In the present research, the Additive Winters model (previous trend analysis) was applied for the double forecast of hospital stays in a public health entity; under an analysis of precision measures, and with a prediction interval of 95%, the forecast model for week 47 and week 48, in this event the values 11 299 and 12 638 stays were predicted with the values MAPE = 6, MAD = 678 and MSD = 649 427, which presents an experimental error for week 47 of 1.46% and for week 48 of 9.23%. To reaffirm the robustness of the model, an analysis of precision measures was conducted, accompanied by residual analysis (Normal probability plot, Histogram, vs fits, and vs order). Similarly, an autocorrelation analysis was performed to reaffirm the model's rigor. The analysis of residuals showed the absence of statistical incidents in both trials. It could be concluded that the Additive Winters model applied in a composite manner presents accurate results in relation to hospital stays, it is important to consider the relevance of the Smoothing Constants analysis to optimize the results.