Prediction of complete lactations is essential for selection decisions and production planning (quota, summer and winter production). A large-scale retrospective study (Weller and Ezra) concluded that daily milk analysis may represent more accurately cow lactation production than periodic analysis. The study objective was to adapt Weller and Ezra's methods for prospective endpoints. Performance to d 305 in lactation was predicted for individual herds and the individual cows comprising them. Such predictions should allow farmers to make decisions regarding selection and production before breeding.Annual calibration models, were constructed by using calving events with complete 305 d production data. The research employed data collected over 2 years from 15 Israeli Holstein herds with 175 to 721 annual calving events and 11,840 to 13,635 kg annual milk production. Cow and production variables (milk, fat, protein and economy corrected milk) were taken from Afilab in-line milk spectrometer and AfiMilk MPC milk meter. Stepwise multiple regression was used to construct models from training data. For validation, the models were applied to predict the consecutive year's performance employing the test data.The correlations between prediction and actual ECM/305 d was (r = 0.688 to r = 0.934) in 54 DIM when applying the model on the consecutive year's validation data set. However, success criteria for decision-making should be the deviation from actual production (−1.8% to 5.4% of ECM) and cows culled by mistake for selection, 1 − specificity (0.0% to 9.1%). Respective values at 34 DIM were −2.7% to 6.6% for deviation from 305 d production of ECM, herd level and 0.0% to 18.9% for erroneous culling. Predictions as early as 34 DIM were superior to those using routine monthly milk tests. Adding genomic predicted transmitting abilities to models may improve prediction based on early lactation data. Rates of erroneous culling for low fat (kg) in Herd 9 dropped from 21.6 to 16.2% and 11.1 to 8.3% at 54 and 84 DIM respectively, when gPTA was employed.The results allow for low risk selection, culling and production planning at 54 DIM or earlier in lactation.