ABSTRACT Herbivorous insects pose a major threat to crop production, with rice suffering significant yield losses due to infestation by the brown planthopper (BPH). To understand the genetic and metabolic basis of BPH resistance in rice, we conducted metabolomic analysis and performed metabolite‐based genome‐wide association studies (mGWAS) on a rice population composed of 168 varieties, which exhibit a wide range of resistance to BPH. Metabolomic analysis revealed a trend of increasing metabolic divergence with increasing resistance levels compared with the susceptible group, with resistant groups maintaining greater metabolic stability after BPH infestation. Furthermore, using these metabolic biomarkers, we constructed a prediction model for BPH resistance and found that biomarkers in non‐infested rice were sufficient to predict BPH resistance. We identified in total 2,738 single‐nucleotide polymorphisms (SNPs) associated with key biomarkers in non‐infested rice and 1,605 SNPs in BPH‐infested rice. Gene Ontology (GO) enrichment analysis revealed that genes associated with biomarkers were enriched in different pathways between non‐infested and BPH‐infested rice. Notably, the SNP rs6_191562334 was significantly associated with the biomarker β ‐damascenone, which correlated positively with rice resistance to BPH and has been shown to inhibit BPH feeding on rice. Knockout of LOC_Os06g17970 increased β ‐damascenone levels and enhanced BPH resistance in rice. Collectively, this integrated approach provided novel insight into the metabolic and genetic mechanisms underlying BPH resistance and facilitated the development of strategies for sustainable control of BPH.