The United States Coast Guard uses pooled time series analysis to develop a ship and aviation fuel requirement forecasting model. Given the volatility of aviation fuel prices and the USAF dependency on foreign oil, alternative fuel sources are a serious consideration and require forecasting models when conducting comparison studies. This research uses the Coast Guard's methodology to develop an Air Force aviation fuel requirements model for the Air Force Cost Analysis Agency (AFCAA). By pooling 1,442 historical consumption time series data points, two regression models are developed that predict aviation fuel requirements in gallons. The remaining 356 randomly excluded data points are then used to validate the two regression models. The research shows that 100 percent of the least squares estimated gallons consumed fell within a 95 percent confidence interval for the single and the sub macro-level models. However, the single and sub macro-level models are fundamentally flawed as both fail the underlying linear regression assumptions of normality, constant variance, and independence. Although the research produces two models that predict aviation fuel requirements well, the application of either the single or sub macro-level models are discourage without proper understanding of the underlying statistics provided.