The extensively growing use of pesticides in the last decades has led to great improvements in agriculture but also to threats for human health. Studying long term effects to assess individual exposures encounters difficulties in exposure assessment. This article summarizes retrospective methods for assessing occupational exposures to pesticides currently used in epidemiological studies. Questionnaires, environmental and biological monitoring constitute direct assessment of exposure, in an individual approach. But the validity of questionnaires is often poor and the metrology rarely reflects past exposures. Some indirect measures have been used together with job exposure matrices. To establish dose-response relationships is important to quantify the risk, but needs accurate exposure assessment based on quantitative indexes.