Background: The safety of waterbirth is contested because of the lack of evidence from randomized trials and conflicting results. This research assessed the feasibility of a prospective study of waterbirth (trial or cohort). Methods: We conducted a prospective cohort study at an Australian maternity hospital. Eligible women with uncomplicated pregnancies at 36 weeks of gestation were recruited and surveyed about their willingness for randomization. The primary midwife assessed waterbirth eligibility and intention on admission in labor, and onset of second stage. Primary outcomes measured feasibility. Intention-to-treat analysis, and per-protocol analysis, compared clinical outcomes of women and their babies who intended waterbirth and nonwaterbirth at onset of second stage. Results: 1260 participants were recruited; 15% (n = 188) agreed to randomization in a future trial. 550 women were analyzed by intention-to-treat analysis: 351 (waterbirth) and 199 (nonwaterbirth). In per-protocol analysis, 14% (n = 48) were excluded. Women in the waterbirth group were less likely to have amniotomy and more likely to have water immersion and physiological third stage. There were no differences in other measures of maternal morbidity. There were no significant differences between groups for serious neonatal morbidity; four cord avulsions occurred in the waterbirth group with none in the landbirth group. An RCT would need approximately 6000 women to be approached at onset of second stage. Conclusions: A randomized trial of waterbirth compared with nonwaterbirth, powered to detect a difference in serious neonatal morbidity, is unlikely to be feasible. A powered prospective study with intention-to-treat analysis at onset of second stage is feasible.