Heating, ventilation and air conditioning (HVAC) systems are equipments used to maintain satisfactory comfort conditions in buildings. Also, energy consumption of ventilated buildings highly depend on the design, performance and control of HVAC systems. In this paper, nonlinear model of a multi-variable HVAC system is considered in which the control inputs are the air and cool water flow rates. Using thermodynamics and heat transfer rules, differential and consequently state space equations of the system are represented. To achieve a good performance, dynamic variables such as output temperature and relative humidity must be controlled. Using input-output feedback linearization, a PI controller is designed. It is shown that by applying the controller, system tracks from one operating point to another with an appropriate specification of time response. In addition, using feedback linearization guarantees robustness of the system against the parametric uncertainties associated with dynamic model.