Software tools able to predict heating and cooling energy demand can effectively support the improvement of energy efficiency in buildings. The latest development of available technologies, such as free cooling and phase change materials, which exploit the building inertia effects, cannot be analyzed through the commonly used steady-state approaches, where the thermal inertia of the building envelope is neglected and monthly averaged climatic data are taken into account. Furthermore, the need to implement innovative regulation criteria for heating and cooling systems and the coupled study of plant and building dynamics push towards the use of dynamic tools with low computational costs. The present paper investigates the simulation of the thermal performance of a benchmark residential building using a self-developed dynamic code implemented in the dedicated software called Building Energy Performance Simulator (BEPS), validated in a previous authors’ work. To investigate the dynamic characteristics of a building in different working conditions, several simulations have been performed for different European localities with different mathematical approaches. In particular, different levels of wall discretization have been considered, highlighting the importance of the inertia of the building envelope. The results show that the use of a simplified description of the entire building leads to good predictions of its energy demand in dynamic conditions with low computational costs. However, only heating demand prediction can be done if the wall thermal capacitance is lumped in a single node, while at least two nodes are needed to correctly predict the building cooling energy demand during the hot season.