The purpose of the current study was twofold: first, to examine adults’ cross-modal matching of emotional prosody with emotional faces, when the emotions represented are both across (e.g., happy vs. sad), and within the same dimension of valence (e.g., sad vs. angry); second to examine 5-year-olds’ cross-modal matching of between-valence emotions. First, I validated the valence and emotion categories of our emotional prosody stimuli with adults (Expt. 1). Next, I presented adults with the validated prosodic stimuli and asked them to choose the face with the matching emotion in both between- (Expt. 2) and within-valence designs (Expt. 3). Adults correctly matched the emotional prosody with the emotional faces in both experiments. In Expt. 4, 5-year-olds successfully matched emotional prosody with happy, surprised, sad, and angry faces. Across all experiments, participants were more accurate on negative than positive trials, suggesting greater sensitivity to negatively-valenced emotions than positively-valenced emotions (i.e., a “negativity bias”). Collectively, these experiments demonstrate that adults will match emotional prosody to emotional faces, even when asked to make within-valence decisions; and that 5-year-olds will match a broader set of emotional prosody and faces types than previously shown in the literature. Taken together, these results offer insight into adults’ and preschoolers’ coordination of emotional cues across modalities and highlight greater sensitivity to negatively-valenced emotions.