Gyrodactylus salaris has been recorded in 46 Norwegian rivers since 1975 and is considered a threat to Atlantic salmon stocks. The primary introductions of G. salaris (primary infected rivers) have been accounted for by specific events, as reported in the literature. The parasite has subsequently dispersed to adjacent localities (secondary infected rivers). The objective of this paper is to address the occurrence of secondary infections by examining the hypothesis of inter-river dispersal of G. salaris. A dispersal model for the secondary river infections via migrating infected fish is proposed. Due to the limited tolerance of G. salaris to salinity, both freshwater inflow to dispersal pathways and dispersal distance were expected to influence the probability of inter-river dispersal. Eighteen rivers were categorised as primary infected rivers, 28 as secondary infected rivers, and 54 as rivers at risk. Four risk factors: the log10 freshwater inflow; the dispersal distance; the time at risk; and the salmon harvest were combined in a multi-variable logistic regression model of the probability of secondary infection. The final multi-variable model included log10 freshwater inflow (Wald chi-square = 9.93) and dispersal distance (Wald chi-square = 6.48). Receiver operating characteristic analyses of the final model supported freshwater inflow as a strong predictor of G. salaris infection status. The strong influence of the freshwater inflow on the probability of secondary infection adds further support to the hypothesis of inter-river dispersal of G. salaris through fjords.