Three-sided matching decision-making (TSMDM) problems are widespread in the real world. In practical TSMDM scenarios, two of the three sides collaborate, and matching objects on each side give their assessments about matching objects on another side according to multi-attributes with heterogeneous relationships. Given its advantages in describing uncertain information and its credibility, Z-mixture number is utilized to represent assessment from matching object, which offers continuous variable or discrete variable leaves matching objects free to choose. Based on this, this paper constructs a multiple attribute TSMDM model with Z-mixture numbers to deal with such situations. First, the prioritized interactive aggregation operator is proposed to aggregate Z-mixture numbers, which takes the prioritization and various interrelationships among the input arguments into account. Second, a weight determination model based on maximal deviation and fuzzy measure is proposed to determine the importance of attribute coalitions. Furthermore, a three-sided stable matching model with cooperation which pursues the maximization of overall satisfaction of matching objects by considering the stable matching conditions is constructed to obtain the stable matching among matching objects. Lastly, a case for the matching of supply, transportation and demand in seaborne coal exchange is presented to validate the rationality and practicability of the proposed method. Compared with the existing studies, the advantages and effectiveness of the proposed method are demonstrated.