This article presents the current results and details of modeling the Tibetan verbal system in the formal grammar and computer ontology. The partially automated model uses an ontological editor to construct semantic classes for verbs based on their syntactic and semantic valencies following the corpus data. The resulting system plays a necessary pragmatic role in automatic syntactic and semantic analysis and disambiguation of Tibetan texts. The research covers a range of problems concerning Tibetan verbal system, such as modeling auxiliary verbs and copulas, verb compounds, verbs with special case government and others.