In this research, 45800 test day records, collected from 12 herds in Tehran province during years 1999 to 2001 by The Iranian Animal Breeding Center (IABC) were used. Data sets included records for milk, fat and protein percentages. Different regression models were fitted to data to estimate daily milk; fat and protein yield for different combinations of milking and sampling time records. Heritability and genetic parameters were estimated using univariate and multivariate Animal Models through DFREML method based on estimated as well as actual 305-day records. Results showed that daily milk and fat yields could be estimated based on night and noon milking time records with considerable accuracy. Heritability of actual and estimated daily milk, fat and protein yield were estimated to be 0.242&0.227, 0.174 & 0.174 and 0.157& 0.141 respectively in repeated record. Repeatability in these traits was 0.585 &0.551, 0.403 &0.406 and 0.487 & 0.466 respectively. Estimated heritability for actual adjusted 305-day milk and fat yield and percent fat records in first lactation were 0.307, 0.204 and 0.338 while for estimated records were 0.301, 0.193 and 0.367 respectively in univariate models. Heritability of milk and fat yield as well as percent fat were 0.245, 0.176, 0.286 and 0.243, 0.175, 0.361 respectively for actual and estimated records in the multivariate. Rank correlation between estimated breeding value using actual and estimated daily yield records of milk, fat and protein were 1, 0.928, and 0.949; and between breeding value using actual and estimated 305 day records for milk and fat yield as well as percent fat were 1, 0.985 and 0.973 respectively. Milk and fat yields can be accurately predicted based on milk records at noon & night milking, and noon milk percent fat, noon & night milking measurements respectively for cows with three milking times per day.