Rule learning (RL) refers to infants' ability to extract high-order, repetition-based rules from a sequence of elements and to generalize them to new items. RL has been demonstrated in both the auditory and the visual modality, but no studies have investigated infants' transfer of learning across these two modalities, a process that is fundamental for the development of many complex cognitive skills. Using a visual habituation procedure within a cross-modal RL task, we tested 7-month-old infants' transfer of learning both from speech to vision (auditory-visual-AV-condition) and from vision to speech (visual-auditory-VA-condition). Results showed a transfer of learning in the AV condition, but only for those infants who were able to efficiently extract the rule during the learning (habituation) phase. In contrast, in the VA condition infants provided no evidence of RL. Overall, this study indicates that 7-month-old infants can transfers high-order rules across modalities with an advantage for transferring from speech to vision, and that this ability is constrained by infants' individual differences in the way they process the to-be-learned rules.