The goal of the thesis is to employ the information available in the Czech verb valency lexicon VerbaLex in the form of complex valency frames (CVF) into the process of syntactic analysis performed by the Synt parser. The discovered valency frames of each verb and its argument structure should be considered for resolving usual parsing ambiguities by a) pruning incorrect parse hypotheses; and b) adjusting probabilities of (im)probable parses. The work will consists of a theoretical description of the problem as well as a practical implementation included in the Synt system.