The aggregation of individual preference into the collective consensus opinion is a key issue in decision-making science. Generally, the aggregation methods are based on the complete preferences from all decision makers on all decision objects. In practice, it is common that individuals can be allowed to provide their incomplete preference information constantly because adequate evaluations are hardly to be given with the limited professional levels or the only time. In this paper, a novel collaborative method is proposed to tackle the group aggregation problem with incomplete preference information by measuring the individual representative degree for collective opinion. The advantages of the proposed approach include, (1) the preference matrix can be applied to compute the consensus and conflict among individual evaluators, (2) the graph is used to construct preferences and aggregate group opinion based on the individual representative degree for collective opinion, (3) the group consensus is obtained to regulate the group decision-making process, and (4) the proposed method can be applied in web decision-making to mine maximal group consensus sequences. The experiment results on synthetic data sets and real data are analyzed and show that the collaborative aggregation method is effective in incomplete preference data.