Individuals repeatedly infected with malaria acquire protection from infection and disease; immunity is thought to be primarily antibody-mediated and directed to blood-stage infection. Merozoite surface proteins involved in the invasion of host erythrocytes are likely targets of protective antibodies. We hypothesized that Papua New Guinean children (n = 206) who acquire high antibody levels to two Plasmodium vivax merozoite proteins, Duffy binding protein region II (PvDBPII) and the 19-kDa C-terminal region of P. vivax merozoite surface protein 1 (PvMSP119), would have a delay in the time to reinfection following treatment to clear all blood-stage malaria infections. Ninety-four percent of the children were reinfected with P. vivax during biweekly follow-ups for 6 months. Since PvDBPII is polymorphic, we examined whether individuals acquired strain-specific immunity to PvDBPII. Children with high antibody levels to a prevalent PvDBPII allele (O) were associated with a delay in the time to reinfection with the same variant of P. vivax by 25% compared to parasites expressing other PvDBPII alleles (age-adjusted hazard ratio, 0.75 [95% confidence interval, 0.56 to 1.00 by Cox regression]) and 39% lower incidence density parasitemia (P = 0.01). Two other prevalent alleles (AH and P) showed a similar trend of 16% and 18% protection, respectively, against parasites with the same PvDBPII allele and reduced incidence density parasitemia. Antibodies directed to PvDBPII PNG-P and -O were both associated with a 21 to 26% reduction in the risk of P. vivax infections with higher levels of parasitemia (>150 parasites/μl), respectively. There was no association with high antibody levels to PvMSP119 and a delay in the time to P. vivax reinfection. Thus, anti-PvDBPII antibodies are associated with strain-specific immunity to P. vivax and support the use of PvDBPII for a vaccine against P. vivax.