Predictive control is a dynamically developing area in automatic control with direct impact in practice. The aim of research is to produce algorithms of predictive control with new features which will fit better to requirements from practice. It will concern mainly the use of new models of processes which will broaden the area of application mainly toward nonlinear systems. In connection with new models new optimization routines will be designed and used. Better applicability in practice will be supported by the design of more adequate approaches to uncertainty modeling and consecutively the creation of robust algorithms. The whole research is motivated by practical applicability and therefore the effort is concentrated on only to obtain new theoretical results but in particular reliable algorithms.