Focusing on workforce-specific skills, this paper investigates how differences in labour protection across countries shape patterns of increasing industrial robots. First, a decision-making model for the workforce-specific skill acquisition and firm robots is constructed based on a Nash game model with incomplete information, and the theoretical relationship between labour protection and the emergence of industrial robots is established from two different perspectives: the specific effect of skills improvement and the robot substitution effect. Subsequently, drawing on cross-country data from between 1993 and 2019, a generalised difference-in-differences (DID) model is constructed and the impact of labour protection on the increase in industrial robots is identified. The results of the research reveal that the impact of labour protection on industrial automation is not significant. However, clear evidence is provided that the adoption of industrial automation is significantly reduced in industries with greater reliance on specific skills, and that labour protection inhibits the substitution of labour by robots through the effect of specific skills. Workforce-specific upskilling also proves to be an important way to mitigate the impact of industrial automation on the labour market. After considering variable measurement bias, endogeneity issues, and the exclusion of other competing hypotheses, the conclusions remain robust. Finally, focusing on the Chinese context, the paper tests the applicability of the theoretical model in China using the Labour Contract Law implementation as a quasi-natural experiment to explore the impact of labour protection on the rise of industrial robots in China at both industrial and corporate levels.