<p>Accurate quantification of fish feeding intensity is crucial for precision feeding in aquaculture because it directly affects feed utilization and farming efficiency. Although multimodal fusion has proven effective for this task, existing methods often overlook response inconsistency and decision conflicts across modalities, thereby limiting the reliability of quantification results. To address this issue, this paper proposes a progressive multimodal interaction network (PMIN) that integrates image, audio, and water-wave data for fish feeding intensity quantification. Specifically, a unified feature extraction framework is constructed to map inputs from different modalities into a structurally consistent feature space, thereby reducing representational discrepancies among modalities. An auxiliary-modality reinforcement primary-modality mechanism is further designed to promote cross-modal information fusion through channel-aware recalibration and dual-stage attention interaction. Furthermore, a decision fusion strategy based on adaptive evidence reasoning is introduced to jointly model the confidence, reliability, and conflicts of modality-specific outputs, thereby improving the stability and robustness of the final decision. Experiments were conducted on a multimodal fish feeding intensity dataset containing 7089 samples. The results demonstrated that PMIN achieved an accuracy of 96.76% while maintaining low computational complexity and model size, and its overall performance surpassed that of both homogeneous and heterogeneous comparison models. Ablation studies, comparative experiments, and real-world application results further validated the effectiveness and superiority of the proposed method. The proposed method can provide reliable support for automated feeding monitoring and precise feeding decision-making in intensive aquaculture systems.</p>