The paper presents a computer program destined to fit experimental data through the nonlinear models : Y= a * X1α * X2β +b ou Y= a * X1α + b. This kind of model is particularly interesting for forestry biomass estimations (where Y = dry weight ; X1 = diameter ; X2 = height of the tree). The program takes into account : - the research of optimal values for both exponents α and β ; - the possibility to weight the residuals of the regression with a power function of the explanatory data X1 α * X2 β ou X1 α as the case may be ; - the calculation of the confidence bounds (level 95 p. 100) for estimated values ; - the calculation of the error when a such an equation is used to predict an estimation of the total biomass (ΣY) of a population. The practical interest of the program stands in its flexibility (conversational program with possibilities of choice and a partial presentation of the important intermediate results on the screen - helping to choose the best option) and in the fact that it is a self sufficient program for the PC.