Procrastination is a common problem among university students, with at least half of the population reporting great difficulties initiating or completing tasks and assignments. Procrastination can have a negative impact on course grades and the ability to achieve a university degree, but can also lead to psychological distress. Cognitive behavior therapy (CBT) is believed to reduce procrastination, but few studies have investigated its effectiveness in a regular clinical setting. The current study explored its effects using a pragmatic randomized controlled trial comparing treatment delivered during 8 weeks as self-guided CBT via the Internet (ICBT) or as group CBT. In total, 92 university students with severe procrastination were included in the study (registered as a clinical trial on Clinicaltrials.gov: NCT02112383). Outcome measures on procrastination, depression, anxiety, and well-being were distributed at pre- and posttreatment as well as 6-month follow-up. An outcome measure of procrastination was administered weekly. Linear mixed and fixed effects models were calculated, along with improvement and deterioration rates. The results showed large within-group effect sizes on procrastination, Cohen's d of 1.29 for ICBT, 95% Confidence Interval (CI) [0.81, 1.74], and d of 1.24 for group CBT, 95% CI [0.76, 1.70], and small to moderate benefits for depression, anxiety, and well-being. In total, 33.7% were regarded as improved at posttreatment and 46.7% at follow-up. No differences between conditions were observed after the treatment period, however, participants in group CBT continued or maintained their improvement at follow-up, while participants in self-guided ICBT showed some signs of deterioration. The findings from the current study suggest that CBT might be an effective treatment for those struggling with severe procrastination, but that a group format may be better for some to sustain their benefits over time and that the clinical significance of the results need to be investigated further.