The way a reader comprehends a text is influenced by a number of factors; and one factor that may play a critical role is the reader’s mood. The extant literature documents that happy and sad moods elicit different processing styles, but the effects on reading comprehension processes, specifically inferential processing, are not completely understood. This study examined the effects happy, sad, and neutral moods had on the generation of bridging and predictive inferences. After a video mood induction, participants read texts that supported the generation of inferences but at varying constraint levels (strong, weak, control). Participants then completed a lexical decision task after each text. Response times for the lexical decision task were shortest for texts that were strongly constrained, followed by weakly constrained, and finally control texts. Accuracy for the lexical decision task was greatest for strongly constrained, followed by weakly constrained, and control texts. There was a marginally significant effect of mood, such that happy-induced readers generated more predictive inferences than sad-induced readers.