<p>The proliferation of the Personal Internet of Things (PIoT) introduces significant ethical challenges, particularly in handling context-sensitive decision-making under uncertainty and dynamic user environments. Conventional Boolean-based ethical models lack the expressiveness required to address such complexities. This paper presents FuzEth, a novel framework for Fuzzy Delphi-Based Ethical Intelligence that augments traditional rule-based logic with probabilistic and fuzzy inference mechanisms. FuzEth leverages the Fuzzy Delphi Method (FDM) to systematically incorporate expert consensus, enabling adaptive calibration of Ethical Operating Principles (EOPs) and dynamic thresholding of context-aware parameters. In parallel, probabilistic reasoning facilitates the quantitative assessment of ethical outcomes by assigning contextual likelihoods to competing ethical alternatives, thus managing ambiguity and partial knowledge. The framework introduces Adaptive Ethics Modes (AEMs), governed by fuzzy membership functions, to dynamically regulate the system’s ethical behavior in response to situational changes. Experimental validation on simulated PIoT environments demonstrates that FuzEth achieves superior ethical decision fidelity, reduced false-positive ethical violations, and increased system adaptability when compared to static ethical models. The results suggest FuzEth as a viable foundation for scalable, ethically aligned PIoT deployments capable of continuous learning and autonomous ethical adjustment in real-time settings.</p>