Background: Sugar-sweetened beverage (SSB) is known to increase risks for many health problems in the general population, but its association with gestational diabetes mellitus (GDM) is rarely discussed. Consumption amount of SSB could be subjected to recall biases and estimation errors, whereas psychological dependence using a standardized assessment tool would provide more persistent and objective measurements. Therefore, we hypothesize that desire of SSB may play a role in developing GDM. Methods: This prospective cohort study recruited 183 pregnant women, who answered self-describing questionnaires designed to depict SSB use behaviors. The desire to drink SSBs was assessed using the modified Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria for substance use disorder (SUD), with questions specific for SSB use. All participants received a 75-gram oral glucose tolerance test during 24th–28th weeks of gestation to screen for GDM. Results: Age, body mass index, and SSB-related SUD DSM-5 items were significant predictive factors of GDM, with odds ratios of 1.112 (95% confidence interval 1.018–1.214), 1.208 (95% confidence interval 1.079–1.353), and 1.338 (95% confidence interval 1.077–1.664), respectively. SUD DSM-5 scores positively correlate with education level (p = 0.046), frequency of dining out (p = 0.028), sedentary lifestyle (p = 0.001), and negatively with water intake amount (p = 0.033). Conclusions: The current study is the first to find a positive association between SSB intake pattern and GDM risk. Specifically, every SSB-related DSM point scored increases GDM risk by 33%, which did not necessarily reflect on the reported SSB consumption amount. To offer useful and specific behavioral advices, decreasing frequencies of dining out, increasing exercises and encouraging plain water intake might be helpful.