Long terms predictions of electrical energy demand (EED) are important for both electric utilities and enterprise of resource management and land planning. The repercussion and need to have adequate EED predictions is highlighted. New scenes imply greater efforts, both considering new variables, and using efficient methods of computational engineering. The results of EED predictions with different methods and the inclusion of consumer price index (CPI) variable for the electrical system of the Canary Islands, of special interest for being an isolated power system, are evaluated. Keywords: Electricity Demand, Long-Term Prediction, Genetic Algorithms, Artificial Neural Networks, Isolated Power System.