This research examines the L2 output of young Korean ESL learners with respect to early child English grammar. From the comparison of the frequency of clause elements between L1 and L2 clause elements with respect to semantic theta roles, both language data showed that the ""agent-action-object"" structure was most prevalent. The dominant clause structures of Korean ESL language data were the SVO and the SVA structure. The frequencies of the V, the VO, and the SV structure in EFL language data were higher than those of Korean ESL utterances. The syntactic development of the Korean ESL language data was formed around the SVO and the SOA structure, whereas the EFL language data did not reveal any specific order of the development. Much more English language input should be provided to the students in natural ways for their acquisition and we should take into consideration the language difficulty and the natural development of English clause structure at the same time when developing EFL textbooks.