Based on its investigated efficiency on first-year learners’ pair conversations, Recording technique (RT) (audio) was further piloted in the speaking class for second-year students within four months to verify its effectiveness in error identification and correction for students’ topic-based individual talks. Research data was collected from the recordings before and after the error correction from both target learners (experimental group) and control group learners. To collect reliable data, the spotted errors were evaluated owing to a set of criteria that involves five language aspects intended for effective individual performance. Also, the improvement was examined and discussed between the researcher and two other teachers of English. This paper presents a brief summary of common speaking errors identified in the second-year students’ performances and their improvements after being corrected with RT, followed by a recommended teaching model in which RT can be effectively used in the classroom environment.