PurposeGrassland degradation is a global ecological issue that inevitably leads to low livestock production efficiency (LPE). Adoption of appropriate technology is an effective way to improve productivity. However, the rate of technology adoption among herders in less developed pastoral areas is low. Therefore, it is critical to improve the level of technology adoption in order to increase LPE.Design/methodology/approachBased on remote sensing data and survey datasets of herder households in China's Qinghai-Xizang Plateau, this paper innovatively constructs a stochastic production frontier model incorporating grassland productivity (i.e. grassland total net primary productivity) to accurately evaluate LPE and uses fractional regression models to determine the impact of technology adoption on LPE.FindingsThe results show that grassland productivity is essential to estimating LPE, and failing to account for it will result in overestimation. Technology adopters have a technical advantage with respect to average LPE (0.596) when compared with non-adopters (0.540), and technology adoption positively contributes to LPE. Furthermore, compared with profit-seeking technology, pro-environmental technology contributes more to improving LPE, and the combined adoption of both technologies leads to a markedly greater enhancement in LPE.Originality/valueFew studies have empirically analyzed the economic benefits of technologies that most smallholders can afford, and few measure LPE considering grassland productivity. This study fills these gaps, and the findings are highly relevant for policies aimed at encouraging technology adoption and facilitating more efficient livestock production.