Weather routing of ships is traditionally done on basis of climatology, e.g. avoiding the higher latitudes with winter storm tracks on North Atlantic. Nowadays it is considered possible to weather route a ship on basis of actual weather forecasting, thereby optimizing on travel time or fuel. This is a complex, dynamic and uncertain decision making process in a Man-Machine relation. The paper describes the influence of various uncertainties in information and information handling on the effectiveness of this modern approach to weather routing. Since the uncertainties considered also include human performance aspects, a qualitative description of the cognitive decision making process is made, and applied to weather routing in its context (decision making aboard a ship). Two experiments are presented to illustrate and indicate the magnitude of the uncertainties. The hypothesis that these uncertainties can be reduced by proper use of weather routing decision support systems will be discussed.