We investigated the migration patterns of hepatitis C virus (HCV) in China. Partial E1 and/or NS5B sequences from 411 volunteer blood donors sampled in 17 provinces and municipalities located in five large regions, the north-northeast, northwest, southwest, central south, and southeast, were characterized. The sequences were classified into eight subtypes (1a, n = 3; 1b, n = 183; 2a, n = 83; 3a, n = 30; 3b, n = 44; 6a, n = 55; 6n, n = 10; 6v, n = 1) and a new subtype candidate. Bayesian evolutionary analysis by sampling trees of the E1 sequences of the five major subtypes revealed distinct migration patterns. Subtype 1b showed four groups: one is prevalent nationwide with possible origins in the north-northeast; two are locally epidemic in the central south and northwest, respectively, and have spread sporadically to other regions; and the fourth one is likely linked to the long-distance dispersion among intravenous drug users from the northwest. Subtype 2a showed two groups: the larger one was mainly restricted to the northwest and seemed to show a trend toward migration via the Silk Road; the smaller one was geographically mixed and may represent descendants of those that spread widely during the contaminated plasma campaign in the 1990s. Subtype 3a exhibited three well-separated geographic groups that may be epidemically unrelated: one showed origins in the northwest, one showed origins in the southwest, and the other showed origins in the central south. In contrast, subtype 3b had a mixture of geographic origins, suggesting migrations from the southwest to the northwest and sporadically to other regions. Structurally resembling the tree for subtype 3a, the tree for subtype 6a showed four groups that may indicate migrations from the central south to southeast, southwest, and northwest. Strikingly, no subtype 6a strain was identified in the north-northeast. With a population of greater than 1.3 billion and a territory of >9.6 million square kilometers, China has a total of 34 provinces and municipalities. In such a vast country, the epidemic history and migration trends of HCV are thought to be unique and complex but variable among regions and are unlikely to be represented by those observed in only one or at best a few provinces and municipalities. However, due to the difficulties in recruiting patients, all previous studies for this purpose have been based only on data from limited regions, and therefore, geographical biases were unavoidable. In this study, such biases were greatly reduced because we utilized samples collected from volunteer blood donors in 17 provinces and municipalities. To our knowledge, this is the first study in which the HCV isolates represented such a large portion of the country, and thus, the results should shed light on the current understanding of HCV molecular epidemiology.