This thesis presents a mathematical model to estimate results of the metal powder production process via gas atomization. The melt breakup process is divided into several steps: Primary breakup (sheet or jet breakup) is described by Linear Stability Analysis with the aim to identify the parameters governing the liquid disintegration process. Apart from gas velocity and melt throughput, these are mainly process parameters such as gas and melt temperatures, gas pressure as well as gas and melt properties. Primary breakup is followed by ligament and primary droplet formation. The possibility of a further breakup into secondary droplets can be estimated by model equations expressing the characteristic times of particle formation – secondary breakup time and spheroidization time compete with acceleration time (relative velocity decay), cooling/solidification time and oxidation time. Fast solidification and oxidation of the particles lower the probability of producing a fine, spherical product. For evaluation, the model equations are applied to three commonly used gas atomization systems: free fall (or open jet), prefilming and confined direct nozzle. In the first case, melt iron is atomized by hot air, for the other two examples, aluminium was chosen to be atomized by cold air and helium. Calculations are compared to atomization trials published by several authors. Despite the limitations of the model (oxidation is not included, only the axial decline of the gas velocity field at the nozzle exit is considered neglecting two-dimensional effects), the calculations are in good agreement with the atomization results, while remaining deviations can be well explained.