Abstract: Abstract：ObjectivesTraditional emergency decision-making methods for carrier-based aircraft support operations overly rely on historical case databases and commanders' experience, facing challenges such as data sparsity and insufficient exploration of the correlation between accident characteristics and sub-solutions. To address these issues, this paper proposes a multi-view graph contrastive learning model that integrates multi-view self-supervised signals and debiased contrastive loss, providing commanders with flexible and efficient emergency response plan recommendations to enhance decision-making efficiency. Methods This study formulates the emergency decision-making problem for carrier-based aircraft support operations as a task of recommending multiple sub-solutions based on the multi-attribute characteristics of emergencies. A multi-view graph contrastive learning model is proposed. First, the model constructs auxiliary views under two perspectives: non-augmented and augmented. On one hand, it builds an event feature collaboration graph and a sub-solution collaboration graph based on the intrinsic relationships between event features and sub-solutions. On the other hand, it generates two augmented auxiliary views by applying approximate singular value decomposition and noise masking to the event feature-sub-solution interaction graph. By fully leveraging the self-supervised information from the interaction graph and multi-view contrastive learning, the model improves recommendation accuracy. Additionally, a debiased contrastive loss function is adopted to mitigate false-negative interference in traditional contrastive loss by correcting negative sample sampling bias, enhancing both the accuracy and robustness of recommendations. Results The experimental results on the simulated dataset demonstrate that the proposed model outperforms both single-perspective graph contrastive learning models and traditional case-based reasoning methods in terms of Recall, Precision, Normalized Discounted Cumulative Gain (NDCG), and Mean Reciprocal Rank (MRR). Specifically, when the number of recommended sub-solutions is 10, the Recall and NDCG of the proposed model show improvements of 1.14% and 2.39%, respectively, compared to the best graph contrastive learning baseline HGCL, and significant enhancements of 64.36% and 43.1%, respectively, over the case-based reasoning method CBR. Conclusions The results validate the effectiveness of the multi-view graph contrastive learning framework in modeling self-supervised information and alleviating interaction data sparsity.