The burden of childhood malaria remains understudied in the Americas, where Plasmodium vivax causes >70% of all infections. We have previously shown that clinical malaria episodes are rare among young Amazonian infants. Nevertheless, participants in our birth study cohort become as vulnerable to clinical malaria as their mothers after the age of 12 months. It remains unclear whether younger infants are less exposed to infection or less susceptible to disease, once infected with malaria parasites. To address this gap, here we use serology to measure the cumulative exposure to blood-stage P. vivax parasites in cohort participants during their first and second years of life and to determine correlates of protection from malarial infection and disease. Participants (n = 435) were enrolled at birth in the city of Cruzeiro do Sul and followed up until the age of 2 years. We found serological evidence of P vivax blood-stage infection, defined as the presence of IgG antibodies to at least two antigens (PvAMA1, PvDBP, or PvMSP119) during follow-up visits at the age of 1 or 2 years of age, in 23 (5.3%) cohort children. However, most (65.2%) children with serologically defined P. vivax infection had no clinical vivax malaria episode notified and treated with antimalarials during their first two years of life. Mixed-effects Poisson regression models, adjusted for sociodemographic variables, local transmission intensity, and malaria in pregnancy, identified prolonged breastfeeding (total breastfeeding duration ≥12 months) as a predictor of decreased risk of blood-stage P. vivax infection over the first two years of life (risk ratio, 0.202, P < 0.001), although not of decreased risk of notified clinical vivax malaria. We conclude that most P. vivax infections in early childhood may remain undiagnosed and untreated in Amazonian Brazil. Prolonged breastfeeding may contribute to protecting infants from P. vivax infection in this low-endemicity region.