This paper compares the forecasting ability of six electricity forecasting models for New Zealand and the Maldives. They are three growth curve models (Logistic, Harvey Logistic and Harvey models), a multiple linear regression model proposed using economic and demographic factors referred to as the Combined model, a growth curve model that uses economic and demographic factors referred to as the Variable Asymptote Logistic (VAL) model and a Box-Jenkins ARIMA model. The models are compared using goodness of fit to the historical consumption data and their forecasting accuracy in the short, medium and long term. The analyses of the six presented forecasting models showed that the best overall model is the Harvey model that generally gave more accurate or comparable forecasts than the more complex ARIMA and Combined models. This shows that the simple Harvey model could play a significant role in forecasting electricity.