In this review, five strategies to estimate mycotoxin exposure of a (sub-)population via food, including data collection, are discussed with the aim to identify the added values and limitations of each strategy for risk assessment of these chemicals. The well-established point estimate, observed individual mean, probabilistic and duplicate diet strategies are addressed, as well as the emerging human biomonitoring strategy. All five exposure assessment strategies allow the estimation of chronic (long-term) exposure to mycotoxins, and, with the exception of the observed individual mean strategy, also acute (short-term) exposure. Methods for data collection, i.e. food consumption surveys, food monitoring studies and total diet studies are discussed. In food monitoring studies, the driving force is often enforcement of legal limits, and, consequently, data are often generated with relatively high limits of quantification and targeted at products suspected to contain mycotoxin levels above these legal limits. Total diet studies provide a solid base for chronic exposure assessments since they provide mycotoxin levels in food based on well-defined samples and including the effect of food preparation. Duplicate diet studies and human biomonitoring studies reveal the actual exposure but often involve a restricted group of human volunteers and a limited time period. Human biomonitoring studies may also include exposure to mycotoxins from other sources than food, and exposure to modified mycotoxins that may not be detected with current analytical methods. Low limits of quantification are required for analytical methods applied for data collection to avoid large uncertainties in the exposure due to high numbers of left censored data, i.e. with levels below the limit of quantification.