Recently, more and more attention has been paid to ship intelligence. However, in the sensor data acquisition network represented by the integrated ship bridge system, sensor data is collected and transmitted point to point, and the data coupling is strong, which is not conducive to the hierarchical utilization of data. To provide more effective data communication methods and flexible support to applications, intelligent integrated platform is needed by modern ship. Prototype design and practice of a ship intelligent platform is proposed, and the key technologies of the platform is discussed. The overall architecture of the platform is described. A hybrid network architecture with fieldbus, real-time ethernet and ethernet information network is introduced. And data storage architecture using NoSQL and hadoop distributed file system is described. The system can meet the real-time performance requirement of the control and information communication. An energy efficiency application based on the designed platform is developed, machine learning based method is employed to predict the heavy oil fuel consumption for ship navigation.