A Monte-Carlo simulation method for the reconstruction of one-dimensional probabilistic mechanical and fatigue strengths is developed to realize reliability analysis at arbitrary probability and confidence levels. This method is valid to the cases of the probabilistic parameters given at special probabilities (Ps) or probability-confidences (P-Cs). To overcome the shortage of common simulations with numerous samples which is out of true production practice, a simulation policy is newly suggested with 7 to 20 samples for material small specimens and, at most 10 for structural specimens and 5 percentages for the error of simulated parameters fitting the original ones. Details of the method in practice are studied for the six possible statistical distributions i.e. normal, lognormal, three-parameter Weibull, two-parameter Weibull, maximum value, and minimum value ones. The reconstructions of the fatigue limits of 10 Chinese engineering materials have indicated the availability and feasibility of the present method.