Improving the long-term performance of deep geothermal reservoirs, as an energy source, can lead to a significant increase in efficiency of heat extractions from these assets. This will assist designers, energy firms, managers, and government decision makers to plan and maintain the use of limited available energy resources and hence enhance key sustainable development goals. Enhanced geothermal reservoirs possess a multi-phase behaviour with complex inter-relationship between several parameters that makes the analysis and design of these systems challenging. Often, this challenge is increased when taking into consideration the optimum use of the available resources and induced costs during both creation and exploitation phases. This research presents a novel design approach developed to achieve efficiency and improved long-term performance in doublet enhanced geothermal systems (EGS). The proposed approach is based on an optimisation procedure using a numerical hybrid methodology integrating a multi-objective genetic algorithm with finite element analysis of fully coupled thermal hydraulic processes of reservoirs. The results of the optimisation process are discussed in comparison with data available from a benchmark case study. The results demonstrate a significant improvement in the long-term performance of EGS reservoir, both in terms of thermal power and costs when optimised using the proposed methodology. (C) 2018 Elsevier Ltd. All rights reserved.