The main object of this paper is to define the role of shipbornc manoeuvring predictors and pre-launch simulators in ship operation and new design, as well as to discuss a few selected details of simulator models and man-machine interface. Evidence is given to show that the manually operated free-sailing scale model will fail as a steering simulator in important cases. Records of helm manoeuvres on board large tankers in harbour approaches reveal a need for predictor assistance. A schematic flow diagram (Fig. 14.2) indicates the way modern computer and control theory will make new use of inputs from ship-model testing and hydrodynamic theory. Quasi-steady computer models are reviewed with respect to size and usefulness, and it is shown that a correct linearization is not always achieved by dropping non-linear terms. The on-line system identification technique will make it possible to use formally simple models in a shipborne predictor. The shore-based simulator for research and development or training requires a more complex model to handle effects of waterway confinements and ambient disturbances.