In this paper a new approach for increasing the overall reliability of a monitoring and decision support system will be explained. The focus is on systems used for ship operator guidance with respect to, say, speed and heading. The basic idea is to convert the given system into a fault tolerant system and to improve multi-sensor data fusion for the particular system. Fault isolation is an important part of the fault tolerant design for in-service monitoring and decision support systems for ships. In the paper, a virtual example of fault isolation will be presented. Several possible faults will be simulated and isolated using residuals and the generalized likelihood ratio (GLR) algorithm. It will be demonstrated that the approach can be used to increase accuracy of sea state estimations employing sensor fusion quality test.