It is known that building performance usually involves different criteria such as the energy consumption, thermal comfort, thermal loads and internal temperatures, which can make its assessment more complex. Thus, this paper aims to describe a decision-making process to improve some criteria related to the energy efficiency of residential buildings considering a multicriteria approach. Computer simulation using EnergyPlus v8.3-2014 was performed to assess the thermal and energy performance of a residential building for four different climates in Brazil (Belem, Campo Grande, Curitiba and Florianopolis). The performance criteria were the degree-hours and energy consumption for air-conditioning (both for heating and cooling) in different decision makers' preference scenarios. The performances of eight constructive systems were evaluated considering uncertainties related to the operational variables of the building, obtained through field survey. The results showed that some systems had statistically equivalent performance considering the range of uncertainty. The system of masonry with double massive brickwork performed better for three climates; for Belem, however, an adequate system highly depends on the preference scenario. Nonetheless, different preference scenarios would indicate different adequate systems for all climates. This study used a method of decision-making which enables the possibility for future analyses of other performance alternatives, design variables, and performance criteria. (C) 2016 Elsevier B.V. All rights reserved.