Code-mixing is a common phenomenon in English as a Foreign Language (EFL) classrooms, particularly among bilingual learners who frequently combine elements from different languages during communication. This study aimed to investigate students’ code-mixing practices and their language attitudes toward the use of code-mixing in EFL learning. An explanatory sequential mixed-method design was employed. The quantitative phase involved 72 eleventh-grade students of SMAS Bequranic Bengkalis who completed a 20-item questionnaire consisting of two dimensions: code-mixing practices and language attitudes. The questionnaire data were analyzed using descriptive statistics, including frequency, percentage, and mean scores. The qualitative phase was conducted through semi-structured interviews with seven purposively selected participants to obtain deeper insights into students’ experiences and perceptions. The interview data were analyzed using thematic analysis. The findings revealed that students demonstrated a relatively high level of code-mixing practices, with an overall mean score indicating that code-mixing was frequently employed during English learning activities. Analysis of the interview data identified three types of code-mixing based on Muysken’s such as insertion, alternation, and congruent lexicalization. Among these types, insertion emerged as the most dominant. The findings also indicated that students used code-mixing to overcome vocabulary limitations, facilitate comprehension, maintain communication, reduce speaking anxiety, and enhance confidence in using English. Overall, code-mixing functioned as a supportive communicative strategy that assisted students in participating more actively in EFL learning contexts.