AbstractIn most cases forest practice in Austria use yield tables to predict the growth of their forests. Common yield tables show the increment of pure even-aged stands which are treated in a way the table developer recommends. The usage of these tables in stands which are either uneven-aged, mixed or treated in another way, may lead to inaccurate predictions. To avoid these problems, forest growth models have been developed. Until now they are not widely used in Austria. One reason may be, that most of the models need some input parameters which are usually not gathered by companies. In this work a basal area increment per hectare model has been developed which is based on the input parameters: diameter at breast height, height to diameter ratio, top height at age 100 years and a selection out of several simple competition indices (growing space, basal area of larger trees, competing basal area, crown cross sectional area, crown competition factor, d/dg, d-dg, basal area and stand density index) which are distance independent. The model parametrization was done with seven different statistical methods (linear regression, linear mixed effect model, resistant linear regression, local polynomial regression, lazy learning model, random forest model and neural network model). By using only few input-parameters it should be possible to parametrize this model for many local areas by using inventory data sets of the specific region. The model works in pure and mixed stands of spruce and beech at the Rosaliengebirge. The observed average diameter increment per 5 years is 18.1 mm for spruce and 21.1 mm for beech. The average difference of the predicted and observed diameter-increment on a validation data-set is 0.3 mm for spruce and -0.3 mm for beech within 5 years and the estimated additional spread caused by the model is +-4.5 mm/5 years for spruce and +-4.0 mm/5 years for beech.