Time series of the number of infected individuals is measured to understand the dynamics of an infectious disease and to evaluate the effectiveness of countermeasures against it. However, in fish diseases, it is often difficult to record the time series of the infected fish owing to i) ambiguous clinical symptoms during the onset of disease and ii) cost of diagnosis. To address this issue, we established a mathematical modeling-based method to infer time series of the number of infected fish by coupling the experimental infection and field data on the outbreak of the disease. This method enables to reconstruct time series of the number of infected fish using mortality data obtained from the experimental infection. The reconstructed time series enables to estimate the transmissibility of fish diseases. We studied Oncorhynchus masou virus (salmonid herpesvirus 2) disease (OMVD) using Rainbow trout (Oncorhynchus mykiss) as a model for analysis and estimated the lethality rate of OMVD infection and the basic reproduction number (R0), i.e., transmissibility of OMVD. Our results revealed that larger fish (≥200 g) showed a higher OMVD lethality rate and R0s when compared with smaller fish (<200 g). This suggests that high mortality among large fish does not inhibit the transmission of pathogen when compared with small fish.