This cross-sectional, analytical study was a comprehensive health assessment focusing on dietary quality of 170 randomly selected elderly respondents in Sharpeville, South Africa. The methods included a sociodemographic, health food frequency questionnaire, 24h-recall questionnaires, and anthropometric and biochemical measurements. The low mean +/- standard deviation (SD) dietary diversity score (3.41 +/- 1.34) and food variety score (4.77 +/- 2.2) compared with poverty parameters confirmed household food insecurity in this community. Although three (n = 99, 58.6%) or two (n = 49, 28.9%) daily meals were mostly consumed, these were mainly carbohydrate-based and nutrient-deficient. The cereal group (2.01 +/- 0.81) had the highest mean food variety score +/- SD, followed by dairy (0.62 +/- 0.53) and flesh foods (0.40 +/- 0.53). When a mean adequacy ratio of 70% was used as a cut-off point for nutrient adequacy, it was found that the food variety score must be eight or higher and the dietary variety score must be at least six. These indicators thus have a high ability to identify those respondents with an inadequate diet but lower ability to identify those respondents with a nutritionally adequate diet. The data further showed a trend that with a higher food variety and dietary diversity, a better mean adequacy ratio is reached for this low-income group of elderly subjects. In conclusion, the results showed that food variety and dietary diversity scores give a fairly good assessment of the adequacy of the diet, and scoring dietary diversity is a significant, yet simple tool to identify elderly persons at risk of food and nutrition insecurity.