As first described in the 1980s, the core food intake model allows a precise assessment of dietary nutrient intake and dietary exposure to contaminants insofar as it reflects the eating habits of a target population and covers the most important foods in terms of consumption, selected nutrient and contaminant contribution. This model has been used to set up the sampling strategy of the second French Total Diet Study (TDS) with the aim of obtaining a realistic panorama of nutrient intakes and contaminant exposure for the whole population, useful for quantitative risk assessment. Data on consumption trends and eating habits from the second French individual food consumption survey (INCA2) as well as data from a 2004 purchase panel of French households (SECODIP) were used to identify the core foods to be sampled. A total of 116 core foods on a national scale and 70 core foods on a regional scale were selected according to (1) the consumption data for adults and children, (2) their consumer rates, and (3) their high contribution to exposure to one or more contaminants of interest. Foods were collected in eight French regions (36 cities) and prepared 'as consumed' to be analysed for their nutritional composition and contamination levels. A total of 20 280 different food products were purchased to make up the 1352 composite samples of core foods to be analysed for additives, environmental contaminants, pesticide residues, trace elements and minerals, mycotoxins and acrylamide. The establishment of such a sampling plan is essential for effective, high-quality monitoring of dietary exposure from a public health point of view.