<p>This paper presents a novel islanding detection technique for hybrid AC/DC microgrids based on a combined index signal generated from distinct methodologies for the AC and DC subsystems. For the DC side, the detection signal is derived by computing the square root of the sum of squared voltage and current values. On the AC side, a sinusoidal-based approach is employed, where the trigonometric identity <i>sin</i>^<i>2</i><i>t</i> + <i>cos</i>^<i>2</i><i>t</i> = <i>1</i> is used to construct the detection signal. These signals are fused through the Shiryaev–Roberts (SR) sequential statistical technique to form a unified index that reliably identifies islanding events. This work is the first to apply the Shiryaev–Roberts framework to islanding detection in power systems, enabling a statistical, sample-by-sample decision process that enhances sensitivity to subtle transients while maintaining robustness against noise and disturbances. The proposed hybrid formulation leverages both AC and DC subsystem characteristics within a single detection structure, thereby reducing dependency on system-specific thresholds and minimizing the non-detection zone (NDZ). Extensive simulations under multiple islanding and non-islanding scenarios confirm the method’s accuracy, resilience, and fast detection performance. The system achieves an islanding detection time of 31.6&#xa0;ms, demonstrating its efficacy for reliable and timely protection in hybrid AC/DC microgrids.</p>