Abstract Background and aims Reduced sleep quality is common in advanced age. Poor sleep quality is associated with adverse outcomes, chiefly cardiovascular, in young and middle-aged subjects, possibly because of its association with metabolic syndrome (MetS). However, the correlates of sleep quality in oldest populations are unknown. We evaluated the association of sleep quality with MetS in a cohort of subjects aged 90+. Methods and results We analysed data of 343 subjects aged 90+ living in the Mugello area (Tuscany, Italy). Quality of sleep was assessed using the Pittsburgh Sleep Quality Assessment Index (PSQI). Good quality of sleep was defined by a PSQI score < 5. MetS was diagnosed according to the National Cholesterol Education Program’s ATP-III criteria; 83 (24%) participants reported good quality of sleep. MetS was diagnosed in 110 (24%) participants. In linear and logistic models, MetS was inversely associated with PSQI score ((B = − 1.04; 95% CI − 2.06 to − .03; P = .044), with increased probability of good sleep quality (OR = 2.52; 95% CI 1.26–5.02; P = .009), and with a PSQI below the median (OR = 2.11; 95% CI 1.11–3.40, P = .022), after adjusting. None of the single components of MetS were associated with PSQI (all P values > .050). However, an increasing number of MetS components was associated with increasing probability of good quality of sleep (P for trend = .002), and of PSQI below the median (P for trend = .007). Generalized Additive Model analysis documented no smoothing function suggestive of nonlinear association between PSQI and MetS. Conclusion Our results confirm a high prevalence of poor sleep quality in oldest age; however, in these subjects, MetS seems to be associated with better sleep quality. Additional larger, dedicated studies are required to confirm our results, and, if so, to identify the subsystems involved and the potential therapeutic implications of such an association.