The purpose of this study was to identify possible subtypes of students with learning disabilities based upon gross motor functions. Subjects in a private school for learning difficulties were divided into a group of students with learning disabilities and a comparison group. Gross motor subtests from the Bruininks-Oseretsky Test of Motor Proficiency (Bruininks, 1978) were administered to both learning-disabled and comparison groups. The four subtypes yielded by the K-means iterative partitioning method demonstrated distinct profiles. Cluster membership was shown to be fairly stable by internal validation techniques. The external validity of the four subtypes was verified by a teacher's ratings of students' physical behaviors. It was recommended that the outcome of type-specific remediation and the longitudinal stability of gross motor subtypes be evaluated.