Abstract: To evaluate the reliability of ship equipment and system more scientifically, a multi-attribute group decision evaluation model with real, interval, and intuitionistic fuzzy numbers as the mixed information carriers, and bidirectional normalized projection and improved TOPSIS technique as the framework is constructed. Firstly, hybrid decision information matrix and weighted evaluation matrix are generated based on the questionnaire and attribute weight vector. Then the closeness of each alternative to the positive and negative ideal solution is evaluated by a bidirectional normalized projection measure, and the priority coefficients are calculated. Finally, the larger the priority coefficient value, the higher the ship equipment reliability. Static and dynamic comparative experimental analysis shows that the proposed method is feasible and superior: it overcomes the deficiencies and shortcomings of distance measures such as Euclidean and Hamming and traditional projection measures; it overcomes the implicit shortcomings of other regularized projection methods proposed in the recent literature and has better dynamic monotonicity; and it has a larger range of values for the preference coefficients, which is conducive to comparing and distinguishing between the reliability of different devices.