Background: The combination of a Chinese herbal medicine (CHM) bath and narrowband ultraviolet B (NB-UVB) improved the efficacy of NB-UVB treatment of psoriasis vulgaris, but bath therapy is inconvenient. Oral CHM plus NB-UVB has been tested in clinical practice. This study aims to evaluate whether adding oral CHM could be beneficial for NB-UVB therapy by a systematic review and meta-analysis. Methods: Nine English and Chinese databases were searched from their inception to April 2014. Randomized controlled trials (RCTs) comparing the combination of orally administered CHM and NB-UVB with that of CHM placebo and NB-UVB or NB-UVB alone for treating psoriasis vulgaris and reporting Psoriasis Area Severity Index (PASI) outcomes were included. A systematic review, meta-analysis, risk of bias assessment and the GRADE assessment were conducted in accordance with Cochrane Collaboration methodology to assess the evidence for efficacy outcome. Data were analyzed in RevMan5.2. Results: Eighteen eligible RCTs (n = 1416) were included for systematic review, and 17 (n = 1342) of them were included in meta-analysis. Risk of bias in terms of blinding was high and so was in publication bias. Quality of evidence was low according the GRADE assessment. PASI-60 or above [risk ratio (RR) = 1.35, 95 % confidence interval (CI) 1.26-1.45, I-2 = 5 %, number needed to treat = 4.27] and PASI-90 or above (RR = 1.71, 95 % CI 1.45-2.01, I-2 = 0 %, number needed to treat = 5.92) were higher in the intervention group. The combination treatment conferred a 24 % benefit of PASI-60 or above (83 vs 59 %, RR = 1.35, 95 % CI 1.26-1.45, P < 0.01). The incidence of NB-UVB-induced adverse events was lower in the intervention group (95/464 vs 123/428, RR = 0.66, 95 % CI 0.46-0.96, P < 0.01). Mild gastrointestinal reactions (2.87 %) and liver function impairments (0.68 %) were reported in the intervention group. No serious adverse events were reported. Conclusion: The orally administrated CHM combined with NB-UVB in treating psoriasis vulgaris showed improved efficacy but quality of evidence was low.