Epidemiologic studies have found lower risks of lung cancer in farmers. However, little is known about the types of agricultural activities concerned. In the Agriculture and Cancer cohort, we assessed the relationship between animal farming and lung cancer by investigating the types of animals, tasks, and timing of exposure. Analyses included 170,834 participants from the Agriculture and Cancer (AGRICAN) cohort in France. Incident lung cancers were identified through linkage with cancer registries from enrollment (2005-2007) to 2011. A Cox model, adjusting for pack-years of cigarette smoking, was used to calculate hazard ratios and 95% confidence intervals. Lung cancer risk was inversely related to duration of exposure to cattle (≥40 years: hazard ratio = 0.60, 95% confidence interval: 0.41, 0.89; P for trend < 0.01) and to horse farming (≥20 years: hazard ratio = 0.64, 95% confidence interval: 0.35, 1.17; P for trend = 0.09), especially for adenocarcinomas, but not with poultry or pig farming. More pronounced decreased risks were reported among individuals who had cared for animals, undertaken milking, and who had been exposed to cattle in infancy. Our study provides strong evidence of an inverse association between lung cancer and cattle and horse farming. Further research is warranted to identify the etiologic protective agents and biological mechanisms.