The airline sector is increasingly using predictive models based on historical demand data towards making decisions in intermediate and long term. This study estimates three different models using data from all the Colombian territory regarding national and international flights like the GDP, employment rates, among others socioeconomics variables. Model structures considered include Holt-Winters and variations of time series models (SARIMA and SARIMAX). For comparing the forecasting performance parameters such as SSE, ESS, RMSE, MAE and MAPE were used. The analysis shows significant gains in model fit when GDP, employment rate and tourism factors were included in the SARIMAX model for national flight predictions, and as for national predictions the SARIMA model had the best performance in terms of the mean absolute percentage error, MAPE.