Abstract This study unfolds the advanced process configuration modification in the evolution of a dual mixed refrigerant (DMR) process for natural gas liquefaction, followed by its optimization through a unique approach i.e., teaching–learning self-study optimization (TLSO). The DMR process is improved by replacing Joule Thomson valves with the isentropic cryogenic turbines. To ensure the maximum possible thermodynamic performance of the retrofitted DMR process, the TLSO paradigm is used and evaluated. The energy, exergy, coefficient of performance, and figure of merit are determined and compared with conventional bench-scale DMR process to find the performance improvement opportunities in the proposed cryogenic turbine-retrofitted DMR process. The performance analysis revealed that the proposed optimal retrofitted DMR process could produce LNG using 28.57% less energy than the base case. The detailed thermodynamic evaluation revealed that the proposed DMR process has 64.68% exergy efficiency, 2.42 coefficient of performance, and 41.6% figure of merit, which are 13.37%, 19%, and 11.9%, higher than the conventional DMR process, respectively. This study would significantly help process engineers overcome the challenges of relating energy efficiency of the LNG plants for both onshore and offshore applications.