Abstract: Comprehensive evaluation problem for decision objects is studied in group decision support systems (GDSS). Based on an analysis of limitations of existing methods, a new approach based on improved Bora voting protocol is proposed, by which the agent can get his preference elicitation. Opinions of decision-maker are collected in a simple way. The qualitative evaluation is translated to a quantitative grade and the constructive coalitional manipulation is avoided. It provides an effective comprehensive evaluation algorithm for scoring and ranking involving multi-user and multi-factors in the GDSS by qualitative evaluation and overcomes the drawbacks of existing systems.