Given their large energy footprint and the availability of building energy management systems, airports are uniquely positioned to take advantage of demand response (DR) programs. Although a baseline, which is the estimation of what the load would have been without load reduction, is essential to assess the performance of DR strategies, however, there has been very little published research on developing baselines for airports. Therefore, the research described in this paper aims to develop baseline models specifically intended for airport facilities. Specifically, we propose piece-wise linear regression models for predicting electricity demand using time-of-week, temperature, and flight schedule information. We hypothesize that flight schedule information would help explain a significant portion of the load after temperature and time-of-week information has been accounted for. However, the result reveals that a model with time-of-week and temperature as input variables and trained over specific seasonal data have the best prediction performance. The number of passengers of departure flight schedule is shown to have a positive relationship with the load, but does not improve the model accuracy significantly.