Interpreting the climate information recorded by carbon isotope composition (S13C) in geological archives aids climatology research and climate prediction. However, this interpretation depends on accurately calibrating modern plant S13C responses to climate change. We here present a long-term calibration conducted at a fixed location adjacent to a meteorological station. By eliminating the effects of geographical variables on C3 plant S13C, as well as errors arising from inaccuracies in meteorological data, this study yields reliable relationships between S13C and climatic factors. This study demonstrates that precipitation is the predominant factor influencing C3 plant S13C, accounting for approximately 65 % of the variance in year-averaged S13C for all C3 plants. The coefficients between year-averaged S13C and annual precipitation, growing season precipitation, and summer precipitation are -0.0039 %o/mm, -0.0048 %o/mm, and -0.0040 %o/mm, respectively. Furthermore, we examined the response pattern of S13C to environmental factors at the species level and observed significant variation in these patterns across different C3 plant species. This indicates that isotope variation of a single species or a limited number of species cannot adequately reflect regional climate information due to microenvironmental influences. Consequently, only S13C-climate factor conversion equations based on community-level data are appropriate for paleoclimate reconstruction. Finally, we used our S13C-precipitation conversion equation to reconstruct the precipitation in the western Loess Plateau during the last glacial period. This study provides a robust constraint of the response pattern of C3 plant S13C to environmental factors and develops a refined isotope-precipitation conversion equation for rainfall reconstruction, promoting the application of S13C in paleoclimate research.