This study investigates additive manufacturing parameters of maraging steel parts using laser powder bed fusion. Laser powder bed fusion can be used to quickly manufacture lightweight and strong parts but requires precise calibration of process parameters such as laser power, scanning speed and layer thickness. Maintaining a good compatibility between these factors is important and the ability to predict the part quality is essential due to this method’s complexity and cost. Maraging steel is known for its high strength, hardness and ductility. During additive manufacturing, maraging steel constantly undergoes a phase change from austenite to martensite, due to the constant cooling and heating cycles caused by layer-by-layer manufacturing. This thesis aims to utilize a phase-changing maraging steel material model, create finite element analyses of laser powder bed fusion and employ direct optimization methods to introduce artificial factors to the analyses to align the finite element model to yield consistent results with the physical tests from literature. Then, metamodel of optimal prognosis from the simulation and experiment data is created. Stochastic optimization methods are discussed, and an evolutionary algorithm is trained with the metamodel of optimal prognosis to predict parameter compatibility and identify optimal manufacturing parameters. This thesis prioritizes deflection caused by the residual effects of metal additive manufacturing as the main failure output of the optimization problem. The findings demonstrated that, despite hardships, near residual effect free maraging steel parts could be additively manufactured by utilizing simulation and optimization methods. A method for increasing the accuracy and efficiency of additive manufacturing of maraging steel is proposed, highlighting the benefits of finite element analysis and stochastic optimization methods.