Abstract The aim of this thesis is to establish a reliable, replicable, and consistent Artificial Intelligent (A.I.) system, capable of predicting accurately the turning tracks of ships. The Artificial Neural Networks (ANN) method has been adapted to solve this problem. The physical and operational data of a ship are described and used as inputs into the system in order to predict the turning manoeuvres. The thesis focuses on both approaches of direct and force models. The ship and controllability data such as underwater hull, rudder and propeller are parameterized and introduced into the system in order to build the direct model for simulating ship manoeuvring motion. The developed method has also been explored in order to include the hydrodynamic forces acting on ships. The initial forces and moments acting on the ship have been described and investigated. The neural system is reinstructed and retuned to solve the prediction problem not only for ensuring more accuracy but also to have deeper insights in the effects of hydrodynamic forces on ship motions, especially the turning manoeuvres. To demonstrate creditability, and confidence in the method used, the results of the program performance were tested against data obtained from ship handling simulators. Parallel Artificial Neural Networks (PANN) is formulated and implemented in MATLAB (instructed, tuned and trained) in order to predict the manoeuvring behaviour of different ship types with variation of sizes, displacements, speeds and rudder angles. Results obtained from the models are compared with the results generated by two different simulators using different ships. The system accuracy and consistency are quantified by the standard deviation. Data optimizers with series models have been established and the emphasis came with higher accuracies and better performance than the general model (direct or force model). The prominence of this application leads to wider applications and better abilities to solve multi-problems, related to the focus theme of the research. The thesis investigates and analyses the diverse results obtained from different prediction models. Thus, it leads to one of the essential points in this research in order to realise the range of the coherence of the applied-based approach. It is essential to study these issues and to analyse the performance of both approaches when applying the environmental conditions in the future to ensure a satisfactory prediction system. The outcome results indicate that the introduced system is capable of thoroughly analysing ship manoeuvring motion and of comparing different ships’ parameters. This system provides the users in advance with the characteristics of transient phases of ship motion. It simulates the real manoeuvring motion before any online training. Further, discussions of recent and future applications are stated in this thesis. The introduced approach proved to be systematic and valid and can take on a variety of forms. System identification techniques, theoretical prediction methods and regression analysis results from other application techniques are also discussed in this thesis.