Autoregressive integrated moving average (ARIMA) was applied to make realtime predictions on the Aedes egg populations in three selected dengue hotspots of Penang, Malaysia. The weekly ovitrap collection was carried out to determine the abundance of Aedes eggs in field population in some selected areas. The ARIMA models were able to estimate actual egg abundance using two criteria. The first criteria is determine the reliability of statistics and the second is to measure the accuracy of forecasting ability of the model equation. The parsimonious model with a lowest order of AR or MA and RMSE value of the forecast for each data set was considered the best. ARIMA (1,0,0), ARIMA (2,0,0) and ARIMA (0,1,1) models were judged to be the best fit for the suburban, urban squatter and urban area data sets respectively. The models were able to forecast the number of eggs within a range of one to eleven weeks. The developed models were able to estimate the egg abundance adequately to permit their use in Aedes control programme in Penang Island. Thus, it can be a useful tool for health officials to improve the management of mosquito control and alert the public to reduce the possibility of dengue outbreaks.