This study was carried out to investigate the importance and relative magnitude of environmental factors for 305-day mills yield by using seven different fixed models constructed with the combination of fixed sets of effects such as herd, year, month, parity, herd × month interaction, arid linear and quadratic regressions for age at calving. Data used for this study included a total of 4,485 sets of 305day lactation records of Holstein cows from six dairy farms participating in the registration project of the Korean Animal Improvement Association from 1969 through 1984. The results obtained from this study are summarized as follows: 1. Use of the fixed model 2 including herd, year, month, and linear and quadratic regressions for age at calving was more appropriate than use of any other model to analyze the importance of environmental factors for 305-day milk yield based on total records. 2. In the analyses of records for each lactation set, use of the fixed model 1 including herd, year, month, and linear regression for age at calving was appropriate for first or second lactation records, whereas use; of the fixed model 5 that quadratic regression for age at calving is added to the fixed model 4 was more appropriate for third and later lactation records. These results indicated that the relative importance of quadratic effects of age at calving was increased at third and later lactations. However, the herd × month interaction effects were not significant for any lactation records. 3. The variance component estimates expressed as percent of the total variance for actual milk yield based on total records under fixed models were: 14-16% for herd, 4-5% for year, 1-2% for month of calving, and less than 1% for herd × month interaction. And the relative sizes of each variance component estimates for each lactation records of milk were similar to those based on total lactation records.