Due to the physiological connection with photosynthesis, sun-induced chlorophyll fluorescence (SIF) provides a promising indicator of vegetation physiological changes caused by environmental stress (e.g. water deficiency). SIF response to crop physiological alterations under water stress is complicated by concurrent non-physiological changes. The non-physiological variation stems from crop structure, leaf optical traits (i.e. pigments, leaf water content, and dry matter), and sun-target geometry. This study aims to disentangle the physiological effect from the non-physiological effect on SIF variations caused by water stress, providing more direct insights into the mechanism of SIF response to stress. We parameterized the radiative transfer model (RTM) SCOPE with top-of-canopy (TOC) reflectance and SIF measurements to decouple the joint effects on TOC SIF in sugar beet. SIF and reflectance data were acquired over irrigated and water-stressed plots using an Unmanned Aerial Vehicle (UAV) on two consecutive days. The non-physiological response was quantified with SCOPE by fitting the model parameters to the TOC reflectance measurements. Subsequently, fluorescence emission yield ($Φ$F) was estimated using SIF measurements to represent the actual physiological status. The results demonstrate that SIF variation caused by water stress both at 687 nm and 760 nm was affected by both the physiological alterations in $Φ$F and the non-physiological changes. At both 687 nm and 760 nm, the non-physiological contribution to SIF variations was lower than the contribution of $Φ$F variation induced by water stress. The lower non-physiological contribution was mainly due to the weak combined effect of the fraction of photosynthetically active radiation absorbed by leaf chlorophyll (fAPARchl) and the fluorescence escape fraction (fesc) on SIF responses. This study provided direct insights into the plant physiological status under water stress and further indicated the ability of the approach of combining RTMs, canopy reflectance, and SIF measurements to support the scalable quantitative use of SIF from the leaf to the ecosystem level.