This paper forms a part of a series of studies aiming to estimate the costs of abortion in Iranian dairy herds. In previous studies we have determined mastitis as a significant risk factor for abortion. In order to provide a more reliable estimation of the costs associated with abortion in Iranian dairy herds, the risk of a cow getting infected with mastitis needs to be included. Data from 6 commercial herds and 32,191 cows was assigned to 3-weeks in milk (3-WIM) records from 1 to 567 d after calving (1st–27th 3-WIM). The effect of herd, parity, calving season, past incidence of abortion, cumulative FCM yield level (CFCML), past incidence of mastitis in previous 3-WIM periods (EMAS), days in milk (DIM) and their significant 2-way interactions on mastitis in current 3-WIM period (MAS) were evaluated using a logistic regression model. Mastitis rate (MR) in studied herds was on average 28.3%. Results show that herd, parity, EMAS, CFCML, calving season, and lactation stage significantly (P < 0.01) influenced the risk of MAS. The risk of MAS increased with lactation number. Cows with EMAS had 4.98 times greater risk of MAS compared to cows with no EMAS. Additionally, cows with medium-level CFCML (i.e. 2, 3, and 4) had a higher risk of MAS compared to cows on level 1 and 5. Calving during the spring significantly (P < 0.01) increased the risk of MAS compared to other seasons. Past incidence of abortion, however, was not significantly associated with MAS, but remained in the final model because of the interaction with other factors. It can be concluded that a risk factor analysis with all significant interactions is more informative than a model without the interactions, especially when making the optimal decision for a cow with given characteristics. Moreover, knowledge on the effect of influential factors on mastitis will be useful when designing mastitis control programs at herd or national level.