Weather and tide variables were used to develop regression models with light trap counts of Aedes sollicitans (Walker) and Aedes taeniorhynchus (Wiedemann) from 1984 to 1995. Rains, tides, and temperatures, both during and preceding the sample period, all showed significant correlations with trap counts. These multiple regression models forecasted general population levels during 1996 and 1997, but not the size of the peaks. Therefore, weather alone can predict general trends but cannot be used to accurately predict population levels of Ae. sollicitans and Ae. taenio-rhynchus.