This study explored a way to help Japanese university students write longer essays while maintaining grammatical accuracy. Participants were three groups of students enrolled in a one-year EFL course in different academic years (N = 111), and the number of words they wrote in 30 minutes and the number of errors made per 100 words were compared. To improve the participants' grammatical accuracy, comprehensive coded feedback (e.g., Bonilla, et al., 2018, 2021; Hartshorn, et al., 2010) and selective metalinguistic explanation (e.g., Bitchener & Knoch, 2010; Sheen, 2007) were provided on the 12 paragraphs/essays they submitted. The first, sixth, and last essays were analyzed to assess their "verb tense" and "mechanical" errors. Regarding the length of writing, the first group kept writing about 150 words, the second group was encouraged to increase the length of writing at their own discretion, and the third group was systematically guided to write longer essays by following a prescribed guideline. The ANOVA results showed that the two groups that wrote longer essays significantly outperformed the short-essay group in the length of writing without sacrificing grammatical accuracy. The correlation analyses produced evidence against a possible trade-off between accuracy and fluency (Lambert & Kormos, 2014; Skehan, 2009).