Excessive dependence on the safety factor of deterministic analysis leads to the conservative design of industrial components, resulting in conspicuous waste of manufacturing resources. The novel strategy for probabilistic assessments for structural integrity is proposed based on the direct method, Linear Matching Method (LMM) framework, a series of structural integrity analysis approaches, where the uncertain design conditions are taken into consideration. To address complicated failure modes under high-temperature, the probabilistic LMM (pLMM) framework consists of six modules, including shakedown analysis, ratcheting analysis, direct steady cycle analysis (DSCA), the unified procedure for fatigue and ratcheting analysis (UPFRA), creep rupture analysis and creep-fatigue analysis. Leveraging the load multiplier calculated by LMM, a general form of performance function is constructed that defines the limit state of different failure modes. Besides, the iteration algorithms of the first-order reliability method (FORM) is formulated to acquire the reliability index and failure probability. With different failure modes discussed, the numerical cases demonstrate the application of the proposed probabilistic structural integrity analysis, with corresponding failure risk and reliability index calculated. The pLMM-based reliability analysis can be introduced into the current structural integrity assessment procedures, which is conducive to replacing traditional safety factors, reducing risks and improving structural robustness.