By utilizing continuous measurements of water fluxes and solar-induced chlorophyll fluorescence (SIF) over the entire growing season, we exploit the potential of broadband SIF in predicting plant transpiration (T) in a temperate forest. After reconstructing the full SIF spectrum from the selected absorption lines and simulations from the SCOPE (Soil Canopy Observation Photochemistry and Energy fluxes) model, linear regression (LR) and Gaussian processes regression (GPR) models are used to analyze the relation between T and combinations of different SIF bands. We find that SIF emissions in the near-infrared spectrum (at 720 nm, 740 nm and 760 nm) are more sensitive to T than SIF emissions in the red spectrum (at 685 nm and 687 nm). While conditions such as light and heat stress decouple the relationship between single-band SIF and T, the combination of different SIF bands allows the retrieval of reliable T estimates even in these conditions. Overall, we find that the use of SIF as a proxy for T yields estimates that are at least as accurate as those from traditional transpiration models such as the Penman-Monteith equation, which are input demanding and complex to apply to in situ and satellite data. Specifically, we find that (1) the SIF-T relationship deteriorates when Photosynthetically Active Radiation (PAR), vapor pressure deficit and air temperature exceed biological optimal thresholds; (2) a high leaf area index exerts a negative impact on the SIF-T correlation due to increasing scattering and (re)absorption of the SIF signal; (3) the SIF-T relationship does not change depending on the observation time during the day; and (4) temporal aggregation to days further enhanced the SIF-T correlations. Altogether, our results provide the first ground-based evidence that SIF emission has potential to be a close predictor of plant transpiration, especially when a combination of different SIF bands is considered.