<p>Non-smooth systems pose a substantial challenge for achieving efficient and accurate long-term dynamic simulations. The field is characterized by a persistent trade-off between the limited accuracy of explicit methods and the computational complexity of implicit schemes. This paper introduces a novel explicit framework that synergistically integrates high-order accuracy, event handling, and energy conservation to address this challenge. Its core innovation is treating event localization error as an independent control variable, using a bisection strategy to bound switching time precision. Furthermore, a one-shot Karush–Kuhn–Tucker projection within the explicit integration simultaneously enforces constraints on position, velocity, and energy, achieving effective boundary protection and precise regulation within a purely explicit scheme. Long-term simulations demonstrate that the method stabilizes the global error at 6 × 10⁻⁴, with event sequences and phase portraits showing excellent agreement with theoretical benchmarks. Consequently, this work enables explicit integration to achieve implicit-level accuracy and stability while retaining its inherent simplicity and parallelism, offering a new pathway for high-fidelity engineering simulations.</p>