In this paper, algebraic difference equation model (Kiviste 2009) for forest growth predictions is examined and adjusted for the use with the data available for Latvian forests. Calculations of soil 0-horizon for Latvian forest types have been made. Linear equation creation from database point approximation has been performed for complete forest stand life cycle coverage. The adjusted algorithm has been implemented to predict Kurzeme region forest situation in the year 2025. Evaluations and updates to the algorithm have been carried out to optimize the results.