Background <p>Medication adherence among patients following total knee arthroplasty is generally suboptimal. Evidence confirms that medication beliefs constitute one of the key factors influencing patients’ adherence to prescribed regimens. However, the mechanism by which medication beliefs affect adherence remains critically important to investigate. Poor medication adherence can significantly compromise postoperative rehabilitation and quality of life in these patients. This study aimed to investigate the mechanism by which medication beliefs influence medication adherence in patients following total knee arthroplasty (TKA).</p> Method <p>This study utilized a moderated mediation analysis. Data were obtained from patients who underwent TKA at four grade-A tertiary hospitals in Shandong Province, China. Upon discharge, all patients were instructed to adhere to regular anticoagulation therapy. Follow-up data were collected during orthopedic outpatient visits 1 month after surgery. The Chinese versions of the Morisky Medication Adherence Questionnaire, Medication Literacy Scale, Medication Beliefs Scale, and the Social Support Scale were employed as measurement instruments. Through moderated mediation and mediation analyses, the relationships among medication beliefs, medication adherence, medication literacy, and social support were examined.</p> Results <p>Medication adherence was significantly higher among patients with greater medication literacy (<i>F</i> = 3.951, <i>p</i> &lt; 0.01). Positive correlations were observed between all measured scales (Medication Adherence Questionnaire, Medication Literacy Scale, Medication Beliefs Scale, and Social Support Scale; <i>r</i> = 0.407, 0.332, 0.206, 0.508, 0.216, 0.244; all <i>p</i> &lt; 0.01). Medication beliefs directly influenced adherence (BootSE = 0.041, 95% CI = 0.077–0.240) and indirectly through medication literacy (BootSE = 0.030, 95% CI = 0.128–0.247), accounting for 45.18% and 54.82% of the total effect, respectively. Social support significantly moderated this mediation pathway (<i>β</i> = 0.124, <i>t</i> = 3.399, <i>p</i> &lt; 0.01), particularly strengthening the relationship between medication literacy and adherence (<i>β</i> = 0.195, <i>t</i> = 5.456, <i>p</i> &lt; 0.001).</p> Conclusion <p>Medication adherence in post-TKA patients is associated with medication beliefs, with medication literacy serving as a partial mediator and social support as a significant moderator. The influence of both medication beliefs and literacy on adherence increases with higher levels of social support.</p>