Abstract: To investigate the metabolic differences in non-volatile compounds across honey processing methods, UPLC-MS/MS-based widely targeted metabolomics was employed to analyze green coffee beans systematically under different mucilage removal rates. The results revealed that 1682 metabolites were identified in green coffee beans subjected to five mucilage removal rate treatments; multivariate statistical analysis demonstrated that the removal rate significantly altered the metabolic profiles of non-volatile compounds. A total of 192 differential metabolites were screened using VIP>1, log₂|FC|>1, and P<0.05 as thresholds. Among them, amino acids and their derivatives, organic acids, phenolic acids, and saccharides emerged as the most dominant metabolite categories, playing pivotal roles in influencing the flavor characteristics of coffee beans. Trend analysis detected 93 non-volatile compounds exhibiting progressively increased relative contents in correlation with decreasing mucilage removal rates. The Kyoto Encyclopedia of Genes and Genomes (KEGG) results indicated that these compounds exhibited a significant enrichment (P<0.05) in inositol phosphate metabolism, pentose and glucuronate interconversions, galactose metabolism, amino sugar and nucleotide sugar metabolism, streptomycin biosynthesis, indole alkaloid biosynthesis, and carbohydrate digestion and absorption. Partial least squares-discriminant analysis (PLS-DA) model showed that ten potential marker compounds predominantly composed of N-acetyl amino acids were distinguished between honey processing methods with varying mucilage removal rates. The findings revealed the regulatory impact of coffee mucilage on the non-volatile compound profiles in coffee beans, providing a theoretical foundation for optimizing coffee flavor attributes through fermentation control strategies.