Some maneuvering experiments are carried out with a ship maneuvering simulator. Six 20, 000 DWT ships are selected as series test ships whose course stability is systematically varied. Firstly, results of computor simulations on automatic course-keeping in steady wind and random sea disturbances are discussed in comparison with the results of real ship measurement. Secondly, results of course-keeping test, path-keeping test and collision avoidance test conducted by captains are discussed emphasizing the relation between course stability of ship itself and maneuvering characteristics of man-machine control system. Principal conclusions obtained are as follows, (1) The increase of rudder area improves maneuverability in sea disturbances much more than the improvement of hull stability. (2) Hysterisis loop width within 2° does not render significant effect on maneuverability, if the rudder area is carefully chosen. (3) Maneuverability should be examined in circumstances that real ships experience as far as possible, because it is firmly affected by sea disturbances, geographical limitations of sea and control system in wheelhouse.