En este trabajo se analiza el error que se comete cuando se pronostican las actividades operativas en una empresa conservera de la zona, utilizando metodos de proyeccion historica cuantitativos: deterministicos (suavizado exponencial y variantes), estocasticos (Box -Jenkins) y de proyeccion enfocada. Estos metodos se emplearon para analizar las ventas ocurridas en 5 anos. Utilizando como criterio de seleccion el metodo de menor RMSE se determino cual era el mejor modelo para los distintos productos. No se pudo obtener un unico metodo de pronostico. Tambien se realizo un analisis acerca de los distintos softwares –Crystal Ball y SPSS- que se utilizaron en la obtencion de los mismos. Posteriormente se mejoraron las proyecciones obtenidas a traves de la alternativa de combinacion lineal de los distintos pronosticos y la alternativa de suavizado exponencial de los datos iniciales. In this work it was analyzed the error that is produced when the operations activities are fore-casted in a fish conserved industry of this region. It was applied several quantitative methods like exponential smoothing and its different versions, Box Jenkins and the forecasting method that it is used in the industry analyzed. The sales in a five years were analyzed . Using as criterion to select the method, the minor error (RMSE) it was determined the best method for every product. Its not found a common method for all products. Moreover, it was analyzed the software used to obtain the forecasting Crystal Ball and SPSS. Better results were found when the results were combined trough of lineal combination and a previous treatment of dates (exponential smoothing) was performed.