ObjectiveTo assess the influence of clinical, psychological, and psychophysical variables on treatment outcomes after application of exercise combined with education with/without manual therapy in people with tinnitus associated with temporomandibular disorder (TMD).MethodsA secondary analysis of a clinical trial was performed investigating the effectiveness of including cervico‐mandibular manual therapy into an exercise combined with education program in 61 subjects with TMD‐related tinnitus. Clinical outcomes including tinnitus severity and tinnitus‐related handicap were assessed at 3 and 6 months post‐intervention. Patients were assessed at baseline for clinical (tinnitus severity, tinnitus‐related handicap, quality of life), physical (range of motion), psychological (depression), and psychophysical (pressure pain thresholds [PPTs]) variables that were included as predictors.ResultsThe regression models indicated that higher scores of tinnitus severity at baseline predicted better outcomes 3 and 6 months post‐intervention (explaining 13% to 41% of the variance) in both groups. Higher scores of tinnitus‐related handicap at baseline predicted better outcome of tinnitus‐related handicap (45% variance) in the manual therapy with exercise/education group. Lower PPTs over the temporalis muscle at baseline predicted poorer clinical outcomes (10.5% to 41% of the variance) in both groups. Other predictors were sex and quality of life (6.7% variance) in the manual therapy group and PPTs over the masseter muscle (5.8% variance) in the exercise/education group.ConclusionThis study found that baseline tinnitus severity and localized PPT over the temporalis muscle were predictive of clinical outcomes in individuals with TMD‐related tinnitus following physical therapy. Other predictors (eg, sex, quality of life) were less influential.