Building energy modeling programs (BEMPs) are effective tools for evaluating the energy savings potential of building technologies and optimizing building design. However, large discrepancies in simulated results from different BEMPs have raised wide concern. Therefore, it is strongly needed to identify, understand, and quantify the main elements that contribute towards the discrepancies in simulation results. ASHRAE Standard 140 provides methods and test cases for building thermal load simulations. This article describes a new process with various methods to look inside and outside the HVAC models of three BEMPs—EnergyPlus, DeST, and DOE-2.1E—and compare them in depth to ascertain their similarities and differences. The article summarizes methodologies, processes, and the main modeling assumptions of the three BEMPs in HVAC calculations. Test cases of energy models are designed to capture and analyze the calculation process in detail. The main findings are: (1) the three BEMPs are capable of simulating conventional HVAC systems, (2) matching user inputs is key to reducing discrepancies in simulation results, (3) different HVAC models can be used and sometimes there is no way to directly map between them, and (4) different HVAC control strategies are often used in different BEMPs, which is a driving factor of some major discrepancies in simulation results from various BEMPs. The findings of this article shed some light on how to compare HVAC calculations and how to control key factors in order to obtain consistent results from various BEMPs. This directly serves building energy modelers and policy makers in selecting BEMPs for building design, retrofit, code development, code compliance, and performance ratings.