Exposure assessment is a poorly understood component of the science of epidemiology. The relationship between exposure to chemicals and ill-health outcomes is often calculated using crude exposure measures such as ever/never exposed or duration. When investigating subtle effects, exposures need to be characterized much more fully in terms of intensity, frequency, duration and route. While occupational exposures tend to be much greater than those experienced from the wider environment there is a need to remember that, for many chemicals, exposure can occur occupationally, environmentally and through consumer use of products containing the material of interest. Inhalation exposure has generally been the traditional focus for most epidemiological investigations but there is now growing awareness of the importance of the dermal and ingested routes of contact and internalization. Quantification of the exposure also needs to be related to a biological mechanism of action and exposure metrics need to be selected accordingly. Occupational exposures can generally be measured using simple well-validated techniques. Environmental exposures require much more sensitive instruments and are more difficult to assess. Exposure modelling, particularly for the environmental fate of chemicals has undergone many recent developments and Monte Carlo techniques can be used to characterize model uncertainty and variability. This approach to exposure assessment can now be used in the setting of the wider environment and will enable a far better understanding of the relationship between exposure and disease.