We examined the effects of a peer-mediated intervention package that taught typically developing children to be more aware of communicative attempts to classmates with disabilities, to use a small set of facilitative strategies ("Stay, play, talk"), and to distribute strategy use across the school day. A multiple baseline design across subjects was instituted with two cohorts of preschool children. Following baseline observations, a total of 8 target children with moderate development disabilities were eventually paired with trained peers who received "buddy training." One trained peer were taught facilitative strategies and encouraged to use them during classroom activities, consistent improvements in social interaction on the part of the trained peers and target children were demonstrated. Similar or more frequent interactions were demonstrated when trained peers were reassigned to different target children during generalization probes. In addition, treatment effects were revealed when comparing sequential analyses applied to the specific communicative behaviors across experimental conditions, in changes in target children's sociometric ratings, and in social validity judgments of videotapes from before and after treatment. This peer intervention approach has promise for improving the communicative interaction and social integration of children with disabilities attending inclusive preschools.