Heating and cooling loads are the major reasons for energy use in buildings. Buildings are usually subject to schedules and set-points which are not optimized in response to the dynamic weather conditions, internal loads, and occupancy patterns. The thermal network model has been widely applied for real-time building load estimation, which is crucial for optimizing the operation of the HVAC system. However, there has been limited exploration of the capabilities of the thermal network model due to constraints imposed by the solution method adopted. In this paper, the exponential matrix method was adopted to simplify the state space equations and solve the thermal network model analytically. This enhances the applications of a simplified thermal network model for investigation of multiple scenarios of HVAC system operations and equipment sizing, and for more accurate estimation of heating and cooling loads. This study also proves that the analytical solution method is asymptotically stable regardless of time step. Atypical office was used as a case study and the predicted building loads are compared with measured data and numerical results from EnergyPlus. For the case study, the model demonstrated better accuracy and is seen to be robust for thermal load estimation for cooling season. (C) 2015 Elsevier B.V. All rights reserved.