This study investigated the limits of electrical energy production from a thermoelectric generator in a high thermal efficiency system using pure water and nanofluids called aluminum oxide and titanium dioxide. It was aim to provide efficient thermal performance by using a Fresnel lens, solar glass tube, heat pipe and tracking system. Experimental measurements were made between 10:00 and 14:00 on July 26 and 27 in 2023. On each measurement day, the system containing pure water and a nanofluid was compared. The values of parameters such as temperature, solar radiation, wind speed and Thermoelectric Generator (TEG) open circuit voltage were measured instantaneously during the measurements. These data were used to calculate thermal resistance, electrical efficiency and output power. The results showed that the open circuit voltage increased with increasing solar radiation and, therefore, hot side temperature. The calculations showed that the highest open circuit voltage of 2.76 V was obtained with the mechanism using pure water as the working fluid. However, the obtained electrical efficiency ηe and output power Pmax values remained at 1.06 % and 1.058 W, respectively. Numerical and Artificial Neural Network (ANN) models are widely used to save time and cost in determining the optimal operating conditions of the TEG. Therefore, in addition to the experimental study, a numerical model consisting of TEG and a cooling system, and an Artificial Neural Network (ANN) model were also developed in this study. The average absolute errors for Computational Fluid Dynamics (CFD) and electrical results were obtained in the numerical model as 3.47 % and 2.97 %, respectively. In addition, the average absolute errors in the ANN model were obtained as 0.6 %, 0.5 % and 0.27 % for water, Al2O3 and TiO2 nanofluids, respectively.