Unsteady-state heat transfer problems require considerable computation time for the numerical solution, due to the high space-time dimensionality of realistic scenarios. This article presents a parallel multi-threading global energy balance method for the thermal evaluation of a room in unsteady state composed of different materials. The parallel numerical method applies the shared memory paradigm using a pool of sub-processes (threads) executing in different processing units (cores) for the thermal evaluation of a room in unsteady state using simulations. The experimental validation compares the proposed parallel method with an analytical solution reported in the literature for a case study considering climate conditions of Mérida City in Yucatan, Mexico. Significant reductions are reported in the execution times and accurate performance results are obtained when using up to 16 computing resources.