Genetic risk factors are believed to combine with environmental exposures and contribute to risk of developing temporomandibular disorder (TMD). In this prospective cohort study, 2,737 people without TMD were assessed for common genetic variation in 358 genes known to contribute to nociceptive pathways, inflammation, and affective distress. During a median follow-up period of 2.8 years, 260 people developed first-onset TMD. Hazard ratios (HRs) were computed as measures of association between 2,924 single nucleotide polymorphisms (SNPs) and TMD incidence. After correction for multiple testing, no single SNP was significantly associated with risk of onset TMD. However, several SNPs exceeded Bonferroni correction for multiple comparison or false discovery rate thresholds (FDR=0.05, 0.1, or 0.2) for association with intermediate phenotypes shown to be predictive of TMD onset. Non-specific orofacial symptoms were associated with voltage-gated sodium channel, type 1 alpha subunit (SCN1A, rs6432860, p=2.77×10−5) and angiotensin-I converting enzyme 2 (ACE2, rs1514280, p=4.86×10−5), global psychological symptoms with prostaglandin-endoperoxide synthase 1 (PTGS1, rs3842803, p=2.79×10−6), stress and negative affectivity with amyloid-β (A4) precursor protein (APP, rs466448, p=4.29×10−5), and heat pain temporal summation with multiple PDZ domain protein (MPDZ, rs10809907, p=3.05×10−5). The use of intermediate phenotypes for complex pain diseases revealed new genetic pathways influencing risk of TMD.