A semi-empirical method for selecting the processing parameters of laser cladding is proposed. This phenomenological approach uses simple mathematical formulae, derived from a statistical analysis of measured data, to relate the laser cladding parameters with the geometric features of the clad track. Given the prescribed clad height and available laser beam power, the proposed method allows to calculate values of the scanning speed and powder feed rate which are required to obtain low dilution, pore free coatings, fusion bonded to the substrate. To illustrate the application of this method, variable powder feed rate laser cladding experiments were carried out with Stellite 6 powder on mild steel substrates. In this technique the laser beam power and radius and the processing speed are kept constant, while the powder feed rate is varied along a single track length according to a specified linear function. The expressions derived from the model allowed to plot the experimental data in a coherent manner, revealing the combined role of the different processing parameters.