Hepatitis C virus (HCV) genotype determination is required in clinical practice to establish the dose and duration of antiviral treatment. Although subtype identification does not impact on current therapy this is changing with new specific inhibitors of HCV enzymes and functions which are becoming available worldwide. These new drugs may yield different antiviral responses and resistance profiles. Accurate classification of HCV genotype and subtype is therefore crucial. An “in-house” method was developed for improving HCV subtyping and the results were compared with a second-generation line probe assay (LiPA) used extensively in Portugal. Phylogenetic analysis was undertaken of the C/E1 and NS5B genomic regions of HCV isolated from 72 prisoners with chronic HCV infection and from reference samples. Although LiPA is considered to be a good method for genotyping, HCV was subtyped in only 47.2% of cases compared with 95.8% of cases by the “in-house” method. Molecular data for both C/E1 and NS5B regions were obtained in 88.9% of the samples. Two out of 23 cases of subtype 1a were misclassified as subtype 1b by LiPA. A putative recombinant like RF1_2k/1b, two potential inter-genotypic recombinants 1b/4a and 3a/4a, and also a potential intra-genotypic recombinant 2q/2k in C/E1 and 2k/2a in NS5B were also identified. The “in-house” method enabled HCV to be subtyped accurately with the detection, in some cases, of recombinant viruses or dual HCV infections. Near full-length genomic analysis to characterize these potential recombinant viruses is planned. J. Med. Virol. 85:815–822, 2013. © 2013 Wiley Periodicals, Inc.