<p>This study aims to assess the reliability index of additively manufactured AlSi10Mg using the Naïve Bayes classifier for classifying the fatigue behavior under the load conditions. Given the difficulties of obtaining fatigue life data from laboratory or field-testing and the lack of loading history data, the corresponding stochastic and machine learning modeling are formulated. The fatigue life obtained experimentally is then simulated using Monte Carlo method, which captured the characteristic variability of the AlSi10Mg material. The Naïve Bayes classifier uses stress amplitude and ultimate tensile strength as the input parameter to classify fatigue failure of the material. The sensitivity of failure probability to stress-strength interaction was shown in the posterior probability surfaces, which clearly showed separability between failure and safe regions. Confusion matrix evaluation was used to quantify the Naïve Bayes classifier’s accuracy and its ability to assess and generalize beyond the experimental data domain. The experimental fatigue data and the induced Monte Carlo samples were used to formulate the FORM limit-state function and compute the reliability index, which provides a quantitative measure of the fatigue safety of AM AlSi10Mg. The Naïve Bayes confusion matrix shows that correct classifications are concentrated within the dominant FORM defined safe and failure regions. FORM yields the reliability index measure of fatigue safety, while the Naïve Bayes classifier captures the nonlinear interaction between stress amplitude and tensile strength through induced probability distributions, thereby providing a complementary interpretation of the material integrity.</p>