A hybrid multiple regression/regression trees/time series modeling was used to relate the total daily punctuality at Frankfurt Airport, Germany, to weather, the traffic flow and the airport system state. The selected modeling approach is applied to the annual, the multi-annual and seasonal data of the years 2001 to 2006. Reduction of the initial variables system dimension is achieved through the application of the Principal Component Analysis method. The portion of the variability that can be explained by the model after correction of autocorrelations in the residuals using autoregressive AR models, is above 60%. A slight year-to-year variability in the number of statistically significant model parameters is detected. Upon identification of the terminal-specific delay impact factors the possibility of the 24 hour punctuality forecast for the interval of one year was discussed.