Exploration a valid way to optimize the thermoelectric performance is one of mainstream in present thermoelectric research community. In this paper, taken γ-graphyne as a potential carbon-based thermoelectric material as an example, we investigate the influence of surface fluctuation disorder on the thermoelectric transport properties by means of a nonequilibrium Green's function method. The results show that the surface fluctuation disorder caused by the substrate roughness could dramatically improve the thermoelectric energy conversion efficiency of γ-graphyne nanoribbons, e.g., the thermoelectric figure of merit could approach 3.0 at room temperature (the length and width are about 55.41 nm and 1.45 nm). Such obvious enhancement mainly stems from the significant suppression of both phononic and electronic thermal conductance, but survived high Seebeck coefficient of pristine nanoribbon. The parameters of surface fluctuation disorder, i.e., fluctuation height and correlation length, on the thermoelectric properties of γ-graphyne nanoribbons are explored as well in this study. The findings presented in this work prove that the surface fluctuation disorder is indeed an efficient approach to boost the thermoelectric properties of γ-graphyne, and provide a guideline to design and fabricate new carbon-based thermoelectric devices.