With the widespread application of smart arbitration systems, the interactive status of humans in arbitration decision-making has been significantly weakened, and human-machine interaction barriers have become increasingly prominent. In particular, when a smart arbitration system generates outcomes in a single step, human-machine interaction barriers substantially compress the interactive experience to which parties should be entitled. The resulting “loss aversion” intensifies the trust gap between humans and machines, leading parties to develop stronger doubts about and resistance to identical objective outcomes. The traditional objective fairness analytical framework centered on “distributive justice, procedural justice” is unable to effectively respond to the trust crisis caused by human-machine interaction barriers, because it fails to recognize that interactive experience is an important vehicle through which the people genuinely feel fairness and justice in every judicial case. Grounded in the genealogical evolution of organizational justice theory, this paper takes “interactional justice” as the core point of entry for overcoming human-machine interaction barriers in smart arbitration. The informational justice and interpersonal justice embodied in interactional justice can overcome human-machine interaction barriers in smart arbitration, reshape the parties’ interactive experience and emotional identification, and thereby effectively enhance human trust in and acceptance of smart arbitration.