We used a DLM that model led the main effects as a linear function and the delayed effects as a polynomial function. The selection of maximum lag was conducted using model residual checking and we checked maximum lag up to 6 months. We used second order quadratic polynomial smoothing for the lag. The mean value of each weather var- iable was used as a baseline (referring value) to measure the relative risks. We plotted relative risks against weather variables and lags to show the entire relationship between weather conditions and dengue.