Objective To address the limitations of existing nature relationship index （NRI） in depicting the human-nature relationship at the individual level in ecologically fragile regions， and their inability to identify regional differences and key constraining factors， this study aims to reveal the overall level， spatial differentiation， major obstacles， and intrinsic trade-offs and synergies of farmers′ relationship with nature in northern Shaanxi. Methods Taking NRI as a theoretical reference， the macro human-nature relationship framework was downscaled to the individual farmer level. An indicator system of the farmers′ relationship with nature index （FRNI） was constructed， consisting of three first-level dimensions， six second-level dimensions， and sixteen third-level indicators. Based on 522 household questionnaires conducted in 25 counties/districts of northern Shaanxi form July to August 2025， index measurement， the obstacle degree model， and Spearman correlation analysis were employed to analyze the overall level， spatial pattern， key obstacles， and trade-offs and synergies of farmers′ relationship with nature. Results 1） The average FRNI in northern Shaanxi was 0.53， indicating a moderate overall level. Among the three first-level dimensions， the "nature is safeguarded" dimension scored the highest （0.61）， the "nature is thriving and accessible" dimension ranked in the middle， and the "nature is used with care and respect" dimension scored the lowest （0.46）. This suggested that although government-led ecological governance and environmental protection have established a certain foundation， farmers remained relatively weak in resource conservation， ecological care， and active participation. 2） The FRNI in northern Shaanxi showed an overall spatial pattern of "higher in the south and lower in the north"， with different dimensions showing differentiated spatial structures. This indicated that regional differences in farmers′ relationships with nature resulted from the combined effects of natural endowment， agricultural production practices， ecological governance efforts， and farmers′ livelihood structures. 3） The main obstacles to improving FRNI were concentrated in limited accessibility to nature， weak water-saving behaviors， and insufficient farmer participation. Overall， relationships among the indicators were dominated by synergies， with a few indicators showing significant trade-offs. Based on obstacle identification and analysis of trade-offs and synergies， a systematic improvement pathway was proposed， which is prioritizing key obstacles， mitigating major trade-offs， and amplifying synergies. Conclusion The perception-based FRNI effectively characterizes the overall status， spatial differences， and key shortcomings of farmers′ relationships with nature in ecologically fragile regions. In northern Shaanxi， improving farmers' relationship with nature depends not only on ecological improvement， but also on the coordinated enhancement of accessibility to nature， water-saving behavior， and farmer participation， with targeted governance tailored to local conditions. The findings provide a support for evaluating farmer-nature relationships， diagnosing critical obstacles， and informing policy in the Loess Plateau and similar ecologically fragile regions.