Abstract: Photosynthetically active radiation (PAR) is a key environmental factor affecting the change of net ecosystem exchange (NEE) during the daytime. However, the coordinate system of PAR measured by horizontal radiometers over sloping terrain does not match that of NEE after tilt-corrected of the ultrasonic anemometer. Using the temperate deciduous broad-leaved forest at the Maoershan site with an average slope of 9° and a azimuth of 296° as a case, we investigated the diurnal variations in NEE and its driving factors in the growing season (May to September) of 2016. We assessed the differences in estimating light response parameters and the explanations of NEE by other environmental factors between the PAR measured by horizontal and slope-parallel radiometers. The results showed that the diurnal change of NEE in each month of the growing season presented a morning-afternoon asymmetrically unimodal curve: the value was negative (net carbon absorption) about 2.5 h after sunrise, reached the peak around 12:00, then approached zero again at two hours before sunset. The daily net uptake maximized in July and minimized in May. During the whole growing season, the time-lag and difference in the PAR measured by the horizontal versus slope-parallel radiometers led to that the PAR values measured by the horizontal radiometer increased photosynthetic quantum yield (α) and daytime respiration rate (R_d) by 13.3% and 11.5%, respectively, and decreased the maximum photosynthetic efficiency (A_max) by 7.7%. The light response curves of NEE were asymmetrical in the morning and afternoon, with R_d and A_max in the afternoon being greater than that in the morning. Weather conditions affected light response parameters: on cloudy days, A_max was higher than that in sunny days, the α and R_d were lower versus those in sunny days for most conditions. However, the monthly A_max and R_d were generally higher for horizontally measured PAR than for slope-parallelly measured PAR, especially for A_max in the cloudy afternoon. The radiometer-orientation affected the explanation of daytime NEE by air temperature (T_a) and vapor pressure deficit (VPD). The correlation of NEE residual based on the slope-parallel radiometer with T_a and VPD (r ranged: 0.013 to 0.197, 0.098 to 0.224) was tighter than that based on the horizontal radiometer (r ranged: 0.082 to 0.219, 0.162 to 0.282) when the negative correlations with T_a for September was excluded. Our results indicated that the measurements of PAR on the inclined terrains could cause errors in the environmental interpretations of NEE. Such findings had implications for the radiometric measurement of mountain vegetation and the reasonable interpretation of carbon exchange in terrestrial ecosystems.