We investigate the role of information frictions in migration. We develop novel moment inequalities to estimate worker preferences while allowing for unobserved worker-specific information sets, migration costs, and location-specific amenities and prices. Using data on internal migration in Brazil, we find that common estimation procedures underestimate the importance of expected wages in migration choices, and that workers face substantial and heterogeneous information frictions. Model specification tests indicate that workers living in regions with higher internet access and larger populations have more precise wage information, and that information precision decreases with distance. Our estimated model predicts that information frictions play a quantitatively important role in reducing migration flows and worker welfare, and limit the welfare gains from reductions in migration costs.