The proportion of cancer attributed to occupation varies, and estimates are partially dependent on the characteristics of exposed populations, type of tumor, and methodology. The main difficulty for estimating these proportions is the reconstruction of past individual occupational exposure. Appropriate methods for retrospective occupational exposure assessment are essential in epidemiological studies in order to avoid misclassification. In this review, qualitative and quantitative methods are discussed, considering that such evaluations require progressive approaches without a previously established hierarchy. Methods to evaluate exposure through environmental and biological measurements, questionnaires, interviews, expert panels, and job-exposure matrices are compared as to their advantages, limitations, accuracy, and validity. In health surveillance, all the above-mentioned methods can be applied, but the use of job-exposure matrices is emphasized, especially using secondary databases.