The adsorption performances of multilayer graphyne-n (n = 1-3) nanostructures (abbreviated as MGN-n) for pure H2S, CH4 and their mixture are explored by the Grand canonical Monte Carlo (GCMC) simulations combined with ideal adsorbed solution theory (LAST). The effects of pressure, temperature, gaseous composition and content of pre-adsorbed water on the adsorption and separation behaviors are examined. For the pure H2S and CH4 adsorption, the dual-site Langmuir-Freundlich (DSLF) model can be used to exactly fit the results of GCMC simulation. Compared with CH4, H2S molecules are preferentially adsorbed in all cases. Due to the effect of adsorbent's structure, the loading of H2S in MGN-1 with smaller free volume (0.662 cm(3) g(-1)) is greater than that in MGN-2 with free volume 0.933 cm(3) g(-1). For the mixture adsorption, the IAST can accurately predict the loading of two species in MGN-1 and MGN-2, but it has a slight deviation for the selectivity predictions. For the MGN-3, due to the strong adsorbate-adsorbate interactions and heterogeneity surface, the LAST can't predict the loading and selectivity precisely. Besides, the separation performances of MGN are compared with zeolites and found that the MGN-1 has the highest H2S adsorption capacity and selectivity, in which the loading and selectivity reaches up to 15.493 mmol g(-1) and 60.8 at 1 MPa and 300 K, respectively. Rising temperature has adverse influence on the H2S/CH4 mixture separation. The H2S selectivity in MGN-1, MGN-2 and MGN-3 reduces 93.51%, 78.24% and 87.46% respectively as the temperature rises from 300 K to 500 K. When the molar fraction of H2S in mixture is low, pre-adsorbed some water is beneficial to the mixture separation. This work demonstrates that the MGN-1 can act as a delightful separation material. (C) 2016 Elsevier Ltd. All rights reserved.