Background <p>Temporal distribution and amplitude of physical activity/inactivity in 24 h known as circadian rest-activity rhythm may predict the risk of various metabolic diseases, including type 2 diabetes (T2D), yet the mechanisms behind the diurnal behavior patterns remain largely unexplored.</p> Methods <p>This study included 74,165 UK Biobank participants who were free of T2D at baseline. Circadian rest-activity rhythm (CRAR) characteristics, including amplitude (strength), acrophase (timing of peak activity), pseudo-F (robustness), and mesor (height), were assessed using an extended cosine model applied to accelerometer data. T2D was assessed using the established UK Biobank algorithms. Using Cox regression and restricted cubic spline models, we examined the association between CRAR and incident T2D as well as subsequent all-cause mortality among individuals developed T2D during the follow-up. Mediation analysis explored the mediating effects of blood and metabolic biomarkers.</p> Results <p>During a median follow-up of 7.9 years, 1784 T2D cases were documented. We found that CRAR abnormalities was associated with a higher risk of incident T2D compared to optimal CRAR, and the multivariate adjusted hazard ratios (HRs) (95% CI) were 1.48 (1.31, 1.67) for low amplitude, 1.25 (1.07, 1.45) for delayed acrophase, 1.17 (1.04, 1.31) for pseudo-F, and 1.55 (1.38, 1.74) for low mesor. Furthermore, low amplitude and low mesor were associated with higher all-cause mortality following the diagnosis of T2D. Serum vitamin D emerged as a crucial mediator in the association between CRAR abnormalities and the risk of T2D as well as subsequent all-cause mortality.</p> Conclusion <p>Our study suggests that CRAR abnormalities are linked to an elevated risk of incident T2D and subsequent mortality. These associations are mediated by blood and metabolic biomarkers, with serum vitamin D playing a significant role as the primary mediator.</p>