An approach to simultaneously estimate parameters of an annual tree growth model was developed, in which the sum of log-likelihood functions for tree survival and diameter growth was maximized. Four methods for acquiring interim values of stand density were evaluated: (1) Updating Attributes, in which individual tree values were summarized at the end of each year within the growth period to predict interim stand-level attributes, (2) Predicting Attributes, in which stand attributes were predicted annually using a stand-level model, (3) Linear Interpolation, in which stand attributes were predicted by linear interpolation, and finally (4) Initial Values, in which stand attributes at the beginning of the growth period were used as predictors throughout the growing period, and the rate of change for tree survival and diameter was assumed to be constant for this period. Data from the Southwide Seed Source Study of loblolly pine (Pinus taeda L.) showed that, overall, the Updating Attributes and Predicting Attributes produced better evaluation statistics in predicting tree survival and diameter growth than did the other two methods. A simulation study confirmed that these two methods produced the least biased estimates of parameters. Compared with the Updating Attributes method, the Predicting Attributes method produced similar evaluation statistics in predicting tree survival and diameter growth and similar bias in estimating model parameters. The Predicting Attributes method therefore offers a reasonable alternative to the Updating Attributes method, because of the ease of programming in available software languages.