This paper proposes an optimal energy planning method for minimizing the cost of heating, ventilation, and air conditioning (HVAC) of a building. Firstly, an RC (Resistance-Capacitance) state space model which can describe the thermal and CO2 dynamics of a building is established. Its parameters can be estimated based on time-series measurements: solar radiation, outside temperature, room temperature, number of occupants, ventilation rate, heating supply, CO2 level, etc. Secondly, this state space model in continuous time domain is rearranged and discretized. Thirdly, the discretized model is converted to constraints and an energy planning method based on linear programming is made for minimizing the costs of CO2 level and room temperature controls. Finally, case studies of a teaching building are carried out to demonstrate the efficacy of the proposed optimal planning method.