This research assessed the predictability of naturalistic evaluation of all-day recordings for speech and language development in children who acquire Korean language. One-day home recordings from 20 Korean children were collected using a Language Environment Analysis (LENA) recorder at 6-8, 12-14, 18-20 months. Both LENA’s automated measures and measures from human coding were obtained from recordings at 6-8 and 12-14 months. A number of different words (NDW), consonant inventory, and utterance structure inventory were identified from recordings of 18-20 months. The results showed that conversation turn count (CTC) and canonical babbling ratio (CBR) at 6-8 months could significantly predict NDW at 18-20 months. Also, CTC and CBR at 12-14 months could significantly predict consonant inventory size at 18-20 months. This study highlights that rather than the amount of speech to children or overheard by children, conversational interactions between infants and caregivers positively affect children’s later speech and language outcomes.