In the present study, during a series of numerical simulations, the process of dimensional optimization of the solar still along with the optimization of the thermoelectric cooling system features connected to it was studied. This method, which is one of the new methods of water production, can be considered as one of the solutions for sustainable water supply to the dominant rural population of Sistan and Baluchestan province located in Iran. Using the Taguchi method as a known method in experimental design, the number of simulations required to perform the optimization process was reduced from 256 modes to 16 modes (16 numerical simulations). The capabilities of the NSGA-II algorithm were also used to find the optimal performance and geometric conditions for solar still. Four effective factors (design variables) were selected in the optimization process, which are cooling rate (Q), LTEC, Hlow, and solar still length, which is indicated by L. By calculating the signal to noise (S/N ratio), the importance of design variables on the quality of solar still performance was determined so that the cooling rate parameter has the highest impact on desalination efficiency and the length of the thermoelectric cooling system as the least important. The L and Hlow parameters were also ranked 2nd and 3rd in importance. Also, signal-to-noise calculations show that increasing the Q parameter (increasing the cooling rate value) alone will lead to an upward trend in desalination production, but this was not observed in other parameters; for example, despite the positive effect of increasing the L on the process of water production has sometimes caused problems with this increase in production, and the opposite is true for the Hlow parameter, and despite the inverse effect of increasing this parameter on the rate of water production, sometimes increasing this parameter leads to increased production. Therefore, it is very important to pay attention to the relative extremes in the mathematical relationship between the design parameters and the target parameter. Also, based on the simulation results of the optimal solar stills introduced by the genetic algorithm, in the right sloping part of the studied geometry, which is the installation site of TEC, the production of fresh water was 20–60% higher than the left part of the geometry. In addition, in the optimal models introduced by the genetic algorithm, the water production efficiency is 25–55% higher than the water production efficiency in the reference solar still studied in the present study.