Abstract: The traditional reliability analysis method of huge data samples cannot be applied to life prediction of the special equipments because of lacking failure samples. Prior information of reliability estimation can be calculated according to the failure data from laboratory during designing and precious records of similar products. Then， the reliability estimation modal was built based on Bayes theorem combining prior information with small samples from field test， and the posterior distribution was obtained. Taking the main beam deflection of overhead travelling crane as one measurable potential failure mode， reliable life prediction of twocycle is realized， i. e. the work time from deflection zero to potential failures and work time from potential failures to functional failure are calculated. The research provides not only some theoretical basis for the main beam inspection and maintenance but also a reference for reliable life prediction and risk prevention of special equipments which lack failure data from factory and laboratory.

Key words: mechanicalelectrical special equipments, reliability assessment, life, crane, Bayes theorem, small sample