OBJECTIVES: To evaluate the cost-effectiveness of biologic agents in the treatment of moderate to severe psoriasis. METHODS: Four biologics were assessed: adalimumab, etanercept, infliximab and ustekinumab. Direct costs were obtained from the perspective of Brazilian Public Health Service (SUS). Probabilities and efficacy were extrapolated from literature.  Unit costs is the Brazilian currency – Real (R$) – and the outcomes considered were PASI90 and PASI75 response (rPASI90 and rPASI75, respectively). For each outcome one Markov Model was created with 12 cycles of 3 months (3 years). Annual discounting of 5% was applied for costs and outcomes. The model considered patients with moderate to severe psoriasis who had Psoriasis Severity Index or Dermatology Quality of Life Index greater than 10 or were unable to take standard systemic therapies. Univariate and probabilistic sensitivity analysis was applied to evaluate parametric uncertainties. RESULTS: Results were expressed as the financial resources demanded for each patient that have achieved and kept rPASI90 or rPASI75. Regarding rPASI90, adalimumab showed to be dominant against the other biologics with a total cost of R$ 64,588.79 (U$ 32,714.78) and 33,7% of patients with rPASI90 after 3 years of treatment (R$ 191,658.13/rPASI90, U$ 97,076.50/rPASI90). When rPASI75 is considered as the measure of outcome, adalimumab and ustekinumab were dominant against infliximab and etanercept, with R$ 124,200.41/rPASI75 (U$ 62,908.58/rPASI75) and R$ 114,230.59/rPASI75 (U$ 57,858.78/rPASI75), respectively. Univariate sensitivity analysis pointed acquisition cost of biologics as the main critical variable. Probabilistic sensitivity analysis shows robustness of our findings.  CONCLUSIONS: From the perspective of SUS, adalimumab is the most cost-effective biological agent for rPASI90. Threshold up to R$ 124,000/rPASI75 (U$ 62,807.07/rPASI75) favors ustekinumab as the most cost-effective drug, while threshold greater than that point to adalimumab. Is being conducted a new Markov modelling to evaluate which sequence of biologic agents is the most cost-effective when failure of initial biological treatment occurs.