The core idea behind graph neural network (GNN)-based remaining useful life (RUL) prediction methods is to obtain effective graph representations, and graph pooling is an efficient means to achieve this. However, existing graph pooling techniques struggle to model hierarchical structures and have limitations in embedding space representation. To address these issues, this paper proposes a hyperbolic hierarchical pooling graph convolutional network (HyPool-GCN) for RUL prediction in multi-source sensor equipment. HyPool-GCN constructs a hierarchical graph pooling framework in hyperbolic space. Using the geometric advantages of hyperbolic space in representing hierarchies, the proposed framework more effectively captures multi-level structural information in graphs, significantly improving the overall structural fidelity of the graph representation. Moreover, this paper proposes an adaptive curvature predictor based on pooling path deviation feedback. By measuring the geometric distortion of paths from leaf to root in hyperbolic space, this predictor dynamically adjusts the curvature parameter, improving the embedding space's adaptability and expressiveness for the graph's hierarchical structure. Finally, performance comparisons on the CMAPSS dataset show that the proposed method outperforms multiple state-of-the-art approaches in prediction accuracy; experiment validation on real-world wind turbine further confirms its practical applicability and potential for broader engineering deployment.