Purpose Considering pain-related beliefs, this study aimed to investigate the influence of disturbed body perception on arm function in patients with frozen shoulder (FS). Methods This study included 90 patients with FS who visited the orthopedic outpatient clinic. We recorded measurements including the Visual Analog Scale (VAS) for pain at rest and during movement, the quick disability of the arm, shoulder, and hand (QDASH) questionnaire for perceived arm function, the short version of the Pain Self-Efficacy Questionnaire-2 (PSEQ-2) for self-efficacy in pain, the short version of the Pain Catastrophizing Scale (PCS-6) for catastrophic thoughts about pain, the short version of the Tampa Scale for Kinesiophobia (TSK-11) for fear of movement, and the Fremantle Shoulder Awareness Questionnaire (FreSHAQ) for disturbed perception around the shoulder. For statistical analysis, we constructed three models. Model 1 used the QDASH score as the dependent variable, with gender, age, duration of disease onset, VAS score at rest, and VAS score during movement as control variables. Model 2 included pain-related beliefs (PSEQ-2, PCS-6, and TSK-11) added to the variables in Model 1. Model 3 further included the FreSHAQ to the variables in Model 2. We calculated the difference in the adjusted coefficients of determination (R² adj) from Model 1 to Model 2 and Model 3 to determine the amount of change (ΔR² adj). Additionally, we calculated the standardized regression coefficient (β-value) of the input factors to examine their effects. Result Hierarchical multiple regression analysis showed a ΔR² adj of 0.13 and 0.17 for Models 2 and 3, respectively, indicating an increase in ΔR²adj after adding the FreSHAQ. Furthermore, the FreSHAQ had the largest effect of all factors, with β = 0.27 (p<0.05). Conclusions The influence of disturbed body perception on perceived arm function in patients with FS was demonstrated, suggesting the importance of considering somatosensory factors in clinical practice.