Research has shown that perceptual disfluencies may affect both actual and predicted memory performance. However, the contribution of perceptual disfluency in multiple modalities to actual and predicted memory has not been investigated and different perceptual modalities may affect these variables to varying extents. The current study investigated how disfluency in visual and auditory modalities may influence actual and predicted memory performance. In a set of three experiments, participants were presented with food recipes in visual and auditory modalities through short clips and were asked to remember these recipes for a later memory test. They also made judgments about the memorability of clips during encoding. The clips were presented in an intact form in visual and auditory modalities, or were distorted in one or both of the modalities. Experiment 1 used a within-subjects design with four study-test cycles, where participants were exposed four complete food recipes. Results revealed that only the distortions in the auditory modality lowered participants’ memory predictions. Experiment 2 used a between-subjects design, in which participants were continually exposed to the same type of perceptual fluency/disfluency condition. This type of design failed to influence memory predictions. For Experiment 3, unique and unrelated steps from different food recipes were selected to eliminate the effect of logical order between items. When the logical order was eliminated, both visual and auditory disfluencies lowered participants’ JOLs, but auditory disfluency affected JOLs more than visual disfluency. Actual memory performance remained unaffected in all three experiments. This study demonstrated that distortions in both modalities jointly affect the JOLs, even though distortions in auditory modality seem to be more effective. The results are discussed in the light of the perceptual fluency hypothesis as well as the use of multiple cues in making memory predictions. When more than one perceptual cue is used, one of the cues might outweigh the other cue under certain conditions.