The building sector is the largest consumer of energy in Morocco with a share of about 36% of the overall energy consumption of the country. This rate is likely to increase if the new buildings are energy consuming. In order to achieve a high energy performance of the building, an optimization of the building envelope is required. This study is often complex and time consuming due to the large number of parameters to consider. The aim of this research work is, on the one hand, to develop and validate metamodels of heating and cooling energy needs for single family houses in six Moroccan climate zones. These metamodels are obtained by regression approach from a set of dynamic simulations of the building behavior. The number of dynamic simulations is reduced drastically by using the design of experiments method. On the other hand, from the metamodels developed, a simultaneous optimization of the building envelope is performed based desirability function concept and the simplex Nelder-Mead algorithm. This methodology is used successfully to perform a fast operational optimization of building envelopes to enhance the implementation of low energy buildings in Morocco. (c) 2015 Elsevier B.V. All rights reserved.