A declarative model consists of semantic constraints which specifies what should be done in a workflow without specifying how it should be done. Consistency checking mechanism for semantic constraints is considered as a crucial step for the verification of workflow models. Undetected inconsistency within the declarative models can cause deadlocks or other sematic errors which cannot be detected from normal control flows of workflow models. In this thesis, we introduce a novel inference mechanism for finding hidden semantic constraints in declarative workflow models. Our experiments have revealed a collection of rules for uncovering hidden semantic constraints.