Traditional key derivation techniques, including the widely adopted PBKDF2, operate with static parameters that do not account for contextual factors such as device capabilities, data sensitivity, or password strength. In this paper, we propose a novel adaptive PBKDF2-based encryption scheme that adjusts its iteration count dynamically based on computational resource index (CRI), data risk level (DRL), and password strength assessment. We present the theoretical model, algorithmic design, and empirical validation of our approach through nine comprehensive experiments, covering performance, scalability, brute-force resistance, entropy quality, and cross-platform consistency. Our results confirm that the adaptive method achieves a secure balance between computational cost and cryptographic strength, outperforming static PBKDF2 in dynamic scenarios. Our framework enhances cryptographic resilience in real-world deployments and offers a forward-compatible foundation for adaptive security solutions.