Existing personalized federated learning methods struggle with non-IID data due to their limited adaptability in client model aggregation. Most approaches either rely on simplistic global-local model combinations or static layer partitioning, failing to capture the complex relationships between clients’ data distributions and model parameters. To overcome these limitations, we propose FedACo, an adaptive federated learning framework that introduces three key innovations: First, a dynamic multi-criteria weighting system that automatically balances parameter similarity, data distribution overlap, and sample size differences during model aggregation. Second, a hierarchical parameter adaptation strategy where early network layers maintain global knowledge while deeper layers progressively specialize for client-specific patterns through intelligent parameter blending. Third, an optimized training objective that jointly preserves model stability and personalization. Experimental results on four benchmarks demonstrate FedACo’s superiority over ten state-of-the-art methods, showing consistent improvements in accuracy. The framework’s design effectively addresses critical limitations in current approaches while maintaining practical deployment feasibility. In addition, the code will be made public after it is accepted.