High-throughput metabolic analysis based on laser desorption/ionization mass spectrometry exhibits broad prospects in the field of large-scale precise medicine, for which the assisted ionization ability of the matrix becomes a determining step. In this work, the gold-decorated hierarchical metal oxide heterojunctions (dubbed Au/HMOHs) are proposed as a matrix for extracting urine metabolic fingerprints (UMFs) of primary nephrotic syndrome (PNS). The hierarchical heterojunctions are simply derived from metal-organic framework (MOF)-on-MOF hybrids, and the native built-in electric field from heterojunctions plus the extra Au decoration provides remarkable ionization efficiency, attaining high-quality UMFs. These UMFs are employed to realize precise diagnosis, subtype classification, and effective prognosis evaluation of PNS by appropriate machine learning, all with 100% accurate ratios. Moreover, a high-confidence marker panel for PNS diagnosis is constructed. Interestingly, all panel metabolite markers present obviously uniform downregulation in PNS compared to healthy controls, shedding light on mechanism exploration and pathway analysis. This work drives the application of metabolomics toward precision medicine.