This paper presents the heat transfer characteristics of a stationary PV system and a dual-axis tracking PV system installed in the Upper Midwest, U.S. Because past solar research has focused on the warmer, sunnier Southwest, a need exists for solar research that focuses on this more-populated and colder Upper Midwest region. Meteorological and PV experimental data were collected and analyzed for the two systems over a one-year period. At solar irradiance levels larger than 120 W/m2, the array temperatures of the dual-axis tracking PV system were found to be lower than those of the stationary system by 1.8 °C, which is a strong evidence of the different heat transfer trends for both systems. The hourly averaged heat transfer coefficients for the experiment year were found to be 20.8 and 29.4 W/m2 °C for the stationary and tracking systems, respectively. The larger heat transfer coefficient of the dual-axis tracking system can be explained by the larger area per unit PV module exposed to the ambient compared to the stationary system. The experimental temperature coefficients for power at a solar irradiance level of 1000 W/m2 were −0.30% and −0.38%/ °C for the stationary and dual-axis tracking systems, respectively. These values are lower than the manufacturer's specified value −0.5/ °C. Simulations suggest that annual conversion efficiencies could potentially be increased by approximately 4.3% and 4.6%, respectively, if they were operated at lower temperatures.