Aim To determine the epidemiological patterns of hepatitis A virus (HAV) infection in urban and rural populations. Methods Study populations were randomly selected by a twostage systematic cluster sampling method. Anti-HAV lgG was defected by enzyme immunoassay. A stochastic process methodtwo-state non-homogeneous Markov chain was used to analyse the seroprevalence of anti-HAV in urban and rural areas of Hebei, China. Two indexes, age-specific Markov risk rate(MRRt) and total Markov risk rate (Total MRR), were introduced. MRRt is a measure of infection risk, and varies as the product of the proportion susceptible at a given age t with the probability of becoming infected in two age/time steps, P^(2)01(t). Total MRR is the weighed sum of MRRts, which provides a reasonable estimate of the disease burden of the whole population. In this paper an ‘elaborated age-grouping method' was firstly raised. Results The total MRR in urban subjects was 109.56, much greater than that (41.16) in rural subjects. When grouping by 10 years of age, MRR1-9. MRR10-19 in urban subjects and MRR1-9 in rural subjects were greater than 20. Grouping by 5 years, the MRR, s of the first four age groups in urban subjects and the first two age groups in rural subjects were greater than 10. Grouping by 2 years, the absolute value of MRRts were big and with an alternative appearance of positive and negative values in urban samples,but the MRR, s above 10 were obviously centralized in 7-16 years of age. in rural subjects, the MRRts had an obvious tendency of decreasing following the increase of age and decreased to less than 10 in age 7-8 years' group. Conclusion The results showed that the disease burden was much heavier in urban populations than that in rural counterparts, and that HA prevention program or HA vaccination program is more eagerly needed in cities than in rural areas. The high risk population of HAV infection are these from 7 to 16 years of age in urban area and these less than 7 years in rural areas, respectively. Since there is