Abstract: Photovoltaic agriculture has been effectively realized for the highly efficient and intensive utilization of clean energy and agricultural space resources. It is of great significance to deploy the photovoltaic power generation facilities and agricultural production activities on the same land space. The land use mode can also promote sustainable development for energy and food security. However, the layout of photovoltaic arrays can inevitably change the original distribution pattern of the solar and thermal environment on the surface, thus forming a significant spatial heterogeneity. The light intensity (shading effect) and duration can then be reduced in the area under the module. There is also a decrease in the spectral components (especially the proportion of photosynthetically active radiation, PAR) in the microclimate factors, such as temperature and humidity. The photothermal parameters can directly dominate the photosynthetic physiological links of the crops, such as the light energy capture efficiency, chlorophyll synthesis, stomatal conductance, and carbon assimilation rate (like Rubisco enzyme activity). These parameters can also interfere with the growth and development rhythm, biomass accumulation, and distribution patterns of the crops. Ultimately, there is a complex and significant impact on the yield potential and quality formation of the crops. Therefore, it is very necessary to explore the heterogeneity of the photothermal environment under photovoltaic arrays and its mechanism on the crop photosynthetic physiology and yield formation. The photovoltaic system can be optimized to screen the suitable crop varieties, and then formulate precise agronomic measures, in order to balance the synergistic benefits of the "photo-electricity-agriculture". This study aims to explore the influence of the different photovoltaic array structures on the internal light environment and the physiological response of the crops. A photovoltaic agricultural system planted with winter wheat was used as the experimental object. A field test was performed on the fixed and tracking brackets of the photovoltaic structures. The optical quantum sensors were used to continuously monitor the solar radiation intensity in the different planting areas (under photovoltaic panels and between plates) during the whole growth period. Simultaneously, the photosynthetic parameters were measured to determine the yield components of the wheat leaves. The test results show that there was significant spatial heterogeneity in the internal light environment of the photovoltaic system. The inter-plate lighting rate in the fixed bracket area reached 84.3%, which was significantly higher than that in the sub-plate area, 37.2% (P<0.05). However, there was no significant difference in the solar radiation intensity between the plates (56.6%) and the subplates (53.4%) in the tracking stent area. The shading effect led to a decrease of 58.7%-61.1% in the average yield of each treatment group, compared with the control. There were the same yield levels of the fixed (2 973.3 kg/hm²) and the traced stent area (3 198.2 kg/hm²). The density of the tracking bracket was similar to that of the fixed one. But the tracking bracket increased the power generation per unit area by 13.2%. The tracking scaffold significantly improved the energy yield efficiency for crop productivity. A promising potential can be expected to coordinate agriculture and photovoltaic cells. This finding can provide support to regulate the optical environment and decision-making on the agricultural photovoltaic projects.