Recent advances in salient object detection struggle with integrating low-level details and high-level semantics, leading to blurred boundaries. Traditional attention mechanisms fail to handle dynamic multi-scale feature interactions, reducing sensitivity to occlusions and small objects. To address these challenges, we propose the Two-stage Feature Aggregation and Multi-scale Attention Network (TFA-MSANet), which includes an Adaptive Attention Module (AAM), Multi-Scale Attention Enhancement (MSAE) module, and Two-Stage Feature Aggregation (TSFA) Decoder. Our method improves feature extraction, aggregation, and cross-scale interaction, leading to more accurate salient object detection. Experimental results on five datasets show its superior performance.