The developed design methodologies both the material and technological ones will make it possible to improve shortly the quality of materials from the light alloys in the technological process, and the automatic process flow correction will make the production cost reduction possible, and - first of all - to reduce the amount of the waste products. Method was developed for analysis of the casting defects images obtained with the X-ray detector analysis of the elements made from the Al-Si-Cu alloys of the AC-AlSi7Cu3Mg type as well as the method for classification of casting defects using the artificial intelligence tools, including the neural networks; the developed method was implemented as software programs for quality control. Castings were analysed in the paper of car engine blocks and heads from the Al-Si-Cu alloys of the AC-AlSi7Cu3Mg type fabricated with the “Cosworth” technological process. The computer system, in which the artificial neural networks as well as the automatic image analysis methods were used makes automatic classification possible of defects occurring in castings from the Al-Si-Cu alloys, assisting and automating in this way the decisions about rejection of castings which do not meet the defined quality requirements, and therefore ensuring simultaneously the repeatability and objectivity of assessment of the metallurgical quality of these alloys.