Background: Despite widespread recognition that the burden of diabetes is rapidly growing in many countries in sub-Saharan Africa, nationally representative estimates of unmet need for diabetes diagnosis and care are in short supply for the region. We use national population-based survey data to quantify diabetes prevalence and met and unmet need for diabetes diagnosis and care in 12 countries in sub-Saharan Africa. We further estimate demographic and economic gradients of met need for diabetes diagnosis and care. Methods: We did a pooled analysis of individual-level data from nationally representative population-based surveys that met the following inclusion criteria: the data were collected during 2005–15; the data were made available at the individual level; a biomarker for diabetes was available in the dataset; and the dataset included information on use of core health services for diabetes diagnosis and care. We first quantified the population in need of diabetes diagnosis and care by estimating the prevalence of diabetes across the surveys; we also quantified the prevalence of overweight and obesity, as a major risk factor for diabetes and an indicator of need for diabetes screening. Second, we determined the level of met need for diabetes diagnosis, preventive counselling, and treatment in both the diabetic and the overweight and obese population. Finally, we did survey fixed-effects regressions to establish the demographic and economic gradients of met need for diabetes diagnosis, counselling, and treatment. Findings: We pooled data from 12 nationally representative population-based surveys in sub-Saharan Africa, representing 38 311 individuals with a biomarker measurement for diabetes. Across the surveys, the median prevalence of diabetes was 5% (range 2–14) and the median prevalence of overweight or obesity was 27% (range 16–68). We estimated seven measures of met need for diabetes-related care across the 12 surveys: (1) percentage of the overweight or obese population who received a blood glucose measurement (median 22% [IQR 11–37]); and percentage of the diabetic population who reported that they (2) had ever received a blood glucose measurement (median 36% [IQR 27–63]); (3) had ever been told that they had diabetes (median 27% [IQR 22–51]); (4) had ever been counselled to lose weight (median 15% [IQR 13–23]); (5) had ever been counselled to exercise (median 15% [IQR 11–30]); (6) were using oral diabetes drugs (median 25% [IQR 18–42]); and (7) were using insulin (median 11% [IQR 6–13]). Compared with those aged 15–39 years, the adjusted odds of met need for diabetes diagnosis (measures 1–3) were 2·22 to 3·53 (40–54 years) and 3·82 to 5·01 (≥55 years) times higher. The adjusted odds of met need for diabetes diagnosis also increased consistently with educational attainment and were between 3·07 and 4·56 higher for the group with 8 years or more of education than for the group with less than 1 year of education. Finally, need for diabetes care was significantly more likely to be met (measures 4–7) in the oldest age and highest educational groups. Interpretation: Diabetes has already reached high levels of prevalence in several countries in sub-Saharan Africa. Large proportions of need for diabetes diagnosis and care in the region remain unmet, but the patterns of unmet need vary widely across the countries in our sample. Novel health policies and programmes are urgently needed to increase awareness of diabetes and to expand coverage of preventive counselling, diagnosis, and linkage to diabetes care. Because the probability of met need for diabetes diagnosis and care consistently increases with age and educational attainment, policy makers should pay particular attention to improved access to diabetes services for young adults and people with low educational attainment. Funding: None.