This study explored the impact of virtual agents and textual enhancement techniques on English language learning. A 2 × 2 mixed-factorial design experiment (N = 60) was conducted using a specially developed VR English learning content, involving two within-subject factors: the roles of virtual agents (teacher vs. clerk) and two between-subject factors: types of textual enhancement techniques (vocabulary only vs. vocabulary with image). The findings revealed no significant interaction between agent types and textual enhancement techniques. However, significant main effects were observed: agent types influenced memory, attitude, and intention to use, while textual enhancements influenced presence. Further implications of the findings were discussed.