<p>Decision-making problems in contemporary world require the collaborative involvement of decision-makers (DMs) from multiple fields. Large-scale group decision-making (LSGDM) integrating diverse perspectives of many DMs has been widely applied across various real-world decision-making scenarios. As the scale of DMs expands, ensuring efficiency while also considering fairness becomes increasingly challenging. To address this, a two-stage consensus-reaching process (CRP) based on minimum adjustment-maximum fairness model is introduced. To balance subgroup cohesion and opinion diversity, we propose a two-phase clustering method incorporating an adaptive adjustment mechanism that promotes the formation of a more balanced subgroup structure. A weight determination method based on preference analysis is proposed to capture differences in DMs’ influence by integrating both multidimensional objective performances and subjective judgments. Furthermore, a two-stage CRP is proposed to promote consensus-reaching at both intra-subgroup and inter-subgroup levels. Specifically, a minimum adjustment-maximum fairness consensus model (MAMFCM) is proposed to ensure adjustment efficiency while better balancing fairness among DMs within subgroups. In the CRP among subgroups, a minimum adjustment consensus model (MACM) with balanced adjustment differences is constructed to generate recommended adjustments for subgroups. A maximum fairness adjustment model (MFAM) is established to guide DMs in adjusting their evaluations, aligning subgroups’ evaluations with recommended adjustments without disrupting internal consensus. Finally, the effectiveness of the proposed model is validated by a numerical example, and its properties are highlighted through comparative analysis. The results indicate that moderate concessions in efficiency to enhance fairness can significantly improve DMs’ satisfaction, thereby facilitating consensus-reaching in LSGDM.</p>