This study was undertaken to examine the association between different patterns of antiseizure medication (ASM) use during pregnancy and adverse obstetric outcomes (preterm birth, low birth weight [LBW], and small for gestational age [SGA]). This retrospective cohort study used the Birth Certificate Application and National Health Insurance data in Taiwan (January 1, 2004 through December 31, 2018). We retrieved weekly ASM among pregnant women with epilepsy who were prepregnancy chronic users and used group-based trajectory modeling to identify distinct patterns of use. Logistic regressions were adopted to examine the association between patterns of ASM use and risk of preterm birth, LBW, and SGA. In addition, we revealed postnatal ASM utilization pattern among these prepregnancy chronic users as an exploratory study. Of 2175 pregnant women with epilepsy, we identified four patterns of ASM use during pregnancy: frequent and continuous (64.87%), frequent but discontinuous (7.08%), intermittent (19.72%), and intermittent and discontinuous users (8.32%). Compared to frequent and continuous users, the adjusted odds ratios for preterm birth in frequent but discontinuous, intermittent, and intermittent and discontinuous users were .83 (95% confidence interval [CI] = .47-1.48), .71 (95% CI = .47-1.05), and .88 (95% CI = .52-1.49), respectively. Similar results were observed for LBW and SGA. In the exploratory study, we found that most of our study subjects maintained the same patterns before and after delivery. After considering duration and timing of exposure, our study did not find an association between four distinct patterns of ASM use and adverse obstetric outcomes among women with epilepsy. The findings suggested that optimal seizure control could be received for pregnant women with epilepsy after evaluating the risks and benefits.