Background We aimed to compare the effectiveness of tofacitinib and ustekinumab in patients with ulcerative colitis (UC) previously exposed to at least one anti-TNF agent. Methods In this multicenter study, we retrospectively included consecutive UC adult patients previously exposed to at least one anti-TNF agent, with partial Mayo score (pMS) > 2, having who started tofacitinib or ustekinumab between January 2019 and June 2022. The primary endpoint was steroid-free clinical remission (pMS ≤ 2) (CFREM) at week 16 (W16). Secondary endpoints were endoscopic remission (CFREM + Mayo endoscopic score (MES) ≤ 1), and histological remission (CFREM + MES ≤ 1 + Nancy index ≤ 1). Comparisons were performed using propensity score analyses adjusted on potential confounders. Results Overall, 289 patients were included (tofacitinib = 124 patients, ustekinumab = 165 patients). The groups were comparable (tofacitinib vs vedolizumab) for male gender (50.8% vs 43.3%), mean age (40.7 vs 42.9 years), median UC duration (8.6 vs 9.3 years), concomitant use of 5-ASA (13.7% vs 9.7%), steroids (25.0% vs 29.7%), immunosuppressants (7.3% vs 5.5%), and ≥ 2 prior biologics (85.5% vs 82.4%). Tofacitinib group had more pancolitis (55.6% vs 42.4%, p = 0.026) and UC with pMS > 6 (64.5% vs 50.3%, p = 0.016). In our study, 42.1 % of the patients treated with tofacitinib continued using a dose of 10 mgx2/day until W16 while 47.3% of the patients on ustekinumab required dose escalation to 90 mg/4 weeks before W16. After propensity score analysis, the rate of CFREM at W16 was 37.8% and 35.6% in the tofacitinib and ustekinumab arms, respectively arm (p=0.75). CFREM at W16 was achieved in 43.3% vs 57.1% (p = 0.48) after failure of one biologic, 20.7% vs 37.9% (p=0.16) two biologics and 46.7% vs 23.2% (p=0.047) or ≥ 3 biologics, in tofacitinib and vedolizumab arms, respectively. After primary failure to at least one biologic, the rate of CFREM at W16 was 46.3% on tofacitinib vs 25.9% on ustekinumab (p = 0.13). CFREM at W16 was similar with tofacitinib and ustekinumab in case of more severe UC such as pMS ≥ 6 (40.6% vs 41.5%) and CRP > 30 (27.2% vs 33.0%). No predictor of tofacitinib effectiveness has been identified. Factors associated with no CFREM at W16 on ustekinumab were male gender (p=0.035), ≥ 3 prior biologics (p=0.013), prior use of tofacitinib (p=0.03), primary failure to at least one biologic (p=0.013). Endoscopic remission was achieved in 17.0% vs 11.7% (p =0.47) and histological remission in 4.4% vs 7.8% (p=0.32) of the patients treated with tofacitinib and ustekinumab, respectively. Conclusion Tofacitinib and ustekinumab have similar effectiveness in UC after anti-TNF failure. However, the efficacy of ustekinumab could be more impacted by prior therapeutic failures.