Using codominant molecular markers (microsatellites) for paternity identification was investigated in hexaploid sweetpotato [Ipomoea batatas (L.) Lam.]. Two experimental populations (CIP and LAES), each consisting of a progeny of known parentage, were scored for the presence or absence of alleles segregating at IB-316 and IB-318 microsatellite loci. Paternity was assessed using paternity exclusion and the most-likely parent methods. In the former, paternity is assigned based on the identification of incompatible parent-progeny marker data. In contrast, the latter method incorporates paternity exclusion and a log-likelihood or LOD score that weighs progeny allelic patterns as to the likelihood that they could have come from a given paternal parent. The number of correctly allocated progeny differed for the methods. Paternity exclusion correctly allocated 7% and 25% of the progeny in the LAES and CIP populations respectively. The most-likely parent method correctly allocated 23% and 88% of the progeny in the LAES and CIP populations, respectively. The greater misassignments in the LAES population were attributed to low allelic diversity at the LAES IB-318 locus and a larger parental population. This study demonstrates the feasibility of identifying paternity in sweetpotato using a minimal number of loci.