Aiming at the problem of mismatch between the prediction model and the actual system due to the performance degradation of the marine diesel engine turbocharging system, an intelligent online prediction model of the diesel engine turbocharging system based on a Bayesian neural network is proposed. The degradation degree of the turbocharging system is characterized by the decay coefficient of the turbocharger efficiency, and the evolution law of the degradation degree of the turbocharging system with the operating time is constructed. Based on the Bayesian neural network, an online update model for the intelligent prediction of turbocharging systems under complex working conditions is established. The uncertainty modeling is introduced through Bayesian inference, combined with the evolution law of the efficiency decay coefficient. The efficiency decay coefficient is updated in real time according to the operating time of the turbocharging system, to realize the high-precision prediction of the pressure and temperature of the turbocharger outlet of the turbocharging system. The proposed method is validated by the simulation data of the diesel engine turbocharging system based on the GT-Power platform. The simulation results show that the average relative errors of turbocharger outlet pressure and temperature of the proposed method are 0.031783% and 0.012129%, respectively, with an overall prediction accuracy of more than 95% under the complex working conditions and different health states of the turbocharging system.