It is important to understand pathogen transmissibility in a population to establish an effective disease prevention policy. The basic reproduction number (R 0) is an epidemiologic parameter for understanding the characterization of disease and its dynamics in a population. We aimed to estimate the R 0 of the highly pathogenic avian influenza (HPAI) subtypes H5N1, H5N8, and H5N6, which were associated with nine outbreaks in Korea between 2003 and 2018, to understand the epidemic transmission of each subtype. According to HPAI outbreak reports of the Animal and Plant Quarantine Agency, we estimated the generation time by calculating the time of infection between confirmed HPAI-positive farms. We constructed exponential growth and maximum likelihood (ML) models to estimate the basic reproduction number, which assumes the number of secondary cases infected by the index case. The Kruskal-Wallis test was used to analyze the epidemic statistics between subtypes. The estimated generation time of H5N1, H5N8, and H5N6 were 4.80 days [95% confidence interval (CI) 4.23-5.38] days, 7.58 (95% CI 6.63-8.46), and 5.09 days (95% CI 4.44-5.74), respectively. A pairwise comparison showed that the generation time of H5N8 was significantly longer than that of the subtype H5N1 (P = 0.04). Based on the ML model, R 0 was estimated as 1.69 (95% CI 1.48-2.39) for subtype H5N1, 1.60 (95%CI 0.97-2.23) for subtype H5N8, and 1.49 (95%CI 0.94-2.04) for subtype H5N6. We concluded that R 0 estimates may be associated with the poultry product system, climate, species specificity based on the HPAI virus subtype, and prevention policy. This study provides an insight on the transmission and dynamics patterns of various subtypes of HPAI occurring worldwide. Furthermore, the results are useful as scientific evidence for establishing a disease control policy.