Objectives: To examine the frequency of street food consumption of people living in low-income settlements in Nairobi and the role of street foods in their daily diet and to reveal why people consume street foods rather than home-prepared foods. Setting, subjects and methods: A cross-sectional descriptive study was done with 1011 households and in-depth interviews with a subsample of 73 households in two selected areas in Nairobi: Korogocho, a low-income slum area and Dandora, a low-middle-income area. Results: The frequency of street food consumption was higher in Korogocho than in Dandora (3.6 vs 2.0 days per week; P<0.001). Street food consumption did not differ between different types of households, with the exception of household size. Employment status of the household head and street food consumption were related (P<0.001): consumption frequency of 3.7 days per week when irregularly or unemployed, 2.9 days/week when self-employed and 2.1 days/week when regularly employed. Furthermore, where an adult woman with primarily a domestic role was present, street food consumption was less (2.55 days per week when present vs 2.95 when not present; P<0.05). Conclusions: Street foods play an important role in the diet of poor households in Nairobi, in particular for breakfast and snacks, because they are cheap and convenient. The frequency of street food consumption is determined by a combination of at least four factors: level of household income; regularity of income; household size; and time available to prepare meals.