Self-reported everyday cognitive failures are common and may interfere with self-management of chronic illness, but evidence from clinical populations in sub-Saharan Africa is limited. We examined associations between cognitive failures and health symptom burden, functional limitation and healthcare utilisation among adults attending healthcare facilities in southwestern Nigeria. Multicentre cross-sectional survey across ten public hospitals in Lagos, Ogun and Oyo states, March–April 2016. The 25-item Cognitive Failures Questionnaire (CFQ; theoretical range 0–100) was the primary outcome (continuous). A study-specific questionnaire captured symptoms (fatigue, shortness of breath, pain; each 0–10), daily activity limitations (4 items × 0–4, range 0–16), healthcare utilisation (counts, past 6 months) and patient-activation confidence (0–12). Primary analysis was multivariable linear regression of continuous CFQ; logistic regression for high CFQ (CFQ > sample cut-point of 32) was a sensitivity analysis. Backward elimination (p > .05 for removal) was applied to an initial model that included age, sex, ethnicity, education, marital status and health indicators; the final reduced model retained health indicators only. Complete-case sample sizes are reported. Of 1,328 eligible attendees, 1,317 consented and 1,299 had a valid CFQ score (M = 31.96 years, SD = 10.97; 62.1% female). In all, 479/1,299 (36.9%) scored above the sample cut-point of 32 (Low CFQ ≤ 32, n = 820; High CFQ > 32, n = 479). In the final linear regression (n = 1,053; R² = .132), higher health-symptom composite (B = 0.65, 95% CI [0.31, 0.99], p < .001), fatigue (B = 0.47, 95% CI [0.07, 0.87], p = .021), daily activity limitations (B = 1.21, 95% CI [0.84, 1.58], p < .001) and recent hospitalisation (B = 2.40, 95% CI [0.68, 4.12], p = .006) were associated with higher CFQ; higher patient-activation confidence was associated with lower CFQ (B = −0.58, 95% CI [−0.90, −0.27], p < .001). The logistic sensitivity analysis (n = 1,053; AUC = .708) was directionally consistent. Greater symptom burden, functional limitation and recent hospitalisation were associated with more frequent self-reported cognitive failures; higher patient-activation confidence was associated with fewer failures after adjustment. The design cannot establish direction, and the CFQ does not measure objective cognitive impairment.