This paper constructs a piece-wise fuzzy linear regression model for terminal delay estimation using the traffic demand and weather. Earlier, the features of traffic demand and weather were used to develop a linear regression model for delay. The accuracy of delay estimation model using a single linear fit is improved by the three-piece fuzzy linear model. Results are presented on the development of the three-piece fuzzy linear model using 360 days data of one terminal area from 2004. The fuzzy methodology, together with the three-step regression model, provides the estimation of the terminal delay performance.