There is potential to optimize lactation lengths on an individual cow basis to allow for extended lactations in cows that exhibit high production and/or persistency at 305 DIM. This requires producers to make delayed breeding decisions within the voluntary waiting period. To identify which cows might be suitable for extended lactation, models can be used to forecast milk yields (MY). In this study, 3 models (Wood, Wilmink and Dijkstra) were fitted to a data set of daily MY (n = 651) from extended lactations >305 DIM. Lactations were grouped by parity (first and greater) and lactation length (305–404 DIM, 405–505 DIM and 505+ DIM). Within each group, the models were fitted to individual curves. Each model had a high goodness of fit for each group, with a mean root mean square prediction error (RMSPE) of 9.3% of the milk yield average. There were no significant difference of goodness of fit between models (P = 0.05) in the parity and lactation length groups, although in each group the Dijkstra model yielded numerically higher fit statistics. Due to this advantage and the mechanistic nature of its parameters, the Dijkstra model was used to predict 305-d MY after fitting to data from only the first i DIM, where i was incremented from 30 to 300 DIM in weekly intervals. Residuals from each 305-d MY prediction were used to calculate RMSPE. In the 30 DIM group, RMSPE was 61136 kg but it declined rapidly to 11.9 kg at 111 DIM, after which additional data only marginally improved the RMSPE to 9.7 kg at 300 DIM. The probability that 305-d MY for an individual cow would be above a given target for inclusion in an extended lactation program was estimable from best-fit parameters and MSPE. In conclusion, the Dijkstra model fitted extended lactations of a variety of lengths in both primiparous and multiparous cows. Furthermore, this model can be fitted to MY data from the first 111 DIM to accurately forecast 305-d MY. These forecasts can be used to identify which individual cows in a herd may be suitable to undergo extended lactation.