Objective To observe the therapeutic efficacy and safety of Chinese herbal fumigation combined with leflunomide (LEF) and prednisone (Pred) in treatment of systemic sclerosis (SSc) complicated pulmonary arterial hypertension (PAH). Methods Totally 99 SSc patients complicated early PAH were randomly assigned to the Western drugs group (WD, 49 cases) and the integrative medicine group (IM, 50 cases). Patients in the WD group took LEF (20 mg) and Pred (15 mg) , once per day. In addition to routine WD program, those in the IM group additionally received Chinese herbal fumigation. All treatment lasted for 6 months. Raynaud's phenomenon (RP) was observed in each group before and after treatment. RP score, erythrocyte sedimentation rate (ESR), C reactive protein (CRP), pulmonary arterial systolic pressure (PASP) , and pulmonary function were compared between the two groups before and after treatment. The clinical efficacy and adverse reactions were evaluated. Results Thirteen cases were lost due to various reasons. A total of 86 patients completed this study, 41 in the WD group and 45 in the IM group. Compared with the same group before treatment, RP score, levels of ESR and CRP all decreased in the two groups after treatment (P <0. 05). Compared with the WM group after treatment, RP score, levels of ESR and CRP were obviously lowered in the IM group after treatment (P < 0. 05). Besides, lowered differences between post-pre-values of ESR, CRP, and PASP were more obviously higher, while elevated differences between post-pre-values of total lung capacity (TLC) and carbon monoxide diffusing capacity (DLCO) were more obviously higher in the IM group (P <0. 05). The total effective rate was 93. 33% (42/45) in the IM group, obviously higher than that in the WD group [70. 73% (29/41) , P <0. 05 ]. There was no statistical difference in total adverse reaction rate between the two groups (x² =0. 019, P =0. 891). Conclusion Chinese herbal fumigation combined with WD had obvious efficacy with less adverse reactions, so it was worth clinical spread.