The traditional order allocation model primarily adopts a multi-objective programming approach. This approach is optimized for profit maximization and cost minimization to obtain the order allocation scheme. However, supplier efficiency is typically an important overlooked index in the order allocation process. In addition, only one decision-maker (DM) usually participates in the decision-making process. Furthermore, the bounded rationality of DMs under the risk state is not considered. Therefore, we established a group game cross-efficiency order allocation model (GGCE-OA) based on regret theory and decision consensus. First, regret theory was used to describe DMs' subjective preferences and capture their bounded rational behaviours. Second, based on the contribution degree and the corresponding iterative algorithm, the weight modification model is established under the group game cross-efficiency evaluation framework to improve the decision consensus among DMs. Third, to allocate orders from the perspective of improving supplier efficiency, we established the GGCE-OA model based on regret theory and decision consensus, as well as establishing the corresponding iterative algorithm. This enabled us to obtain a fair, unique and accepted order allocation scheme by all DMs. Finally, the model established in this paper was applied for order allocations to green suppliers to verify its effectiveness and practicability.