Maintaining shakedown during the design stage is crucial for the engineering components under cyclic loading conditions. Due to the multiplicity of uncertainty existing in the actual operating parameters, probabilistic shakedown analysis is more conducive to dealing with risk management for critical infrastructures, compared with the traditional experience-based safety factors which mainly rely on deterministic evaluations and lack statistical information for decision-making. Aimed at reasonably considering the redundancy in structural integrity assessment against cyclic load conditions, a novel direct method-based probabilistic shakedown analysis is proposed under the new probabilistic Linear Matching Method framework (pLMM). The risk of losing the shakedown state is predicted by the physics-based estimation model, where the efficient iteration is employed to derive the reliability index. In the benchmark, the probabilistic shakedown boundary is constructed, with the additional influence of the uncertain cyclic loading pattern on the reliability fully reflected. This pLMM shakedown analysis facilitates the progress from over-reliance on the conservative safety factor towards precise risk management.