Since the 1990s, grammatical complexity has received substantial research attention in applied linguistics (Bulté and Housen, 2014). The representation of grammatical complexity has expanded in L2 writing with the application of diverse measures in empirical studies in the recent three decades (1991–2020). In response to this situation, we found it important to revisit grammatical complexity, and an exploratory factor analysis was applied to explore latent dimensions (i.e., factors) of grammatical complexity in L2 writing. We analyzed Lu’s (2011) 14 grammatical complexity measures in the L2 corpus of the British Academic Written English Corpus. We then proposed a four-factor model with “clausal subordination,” “phrasal construction”, “global length unit” and “others.” The four factors generally align with the types of grammatical complexity proposed in Lu (2011), but differences on six measures are also found. Noteworthy points were discussed to interpret the reasons behind the differences. Research implications are provided to show further research directions.