The effect of measurement interval on the reliability of parameters estimated by the result of a full scale measurement of responses of ship to ocean waves is an important factor to consider in planning the measurement on a ship. To clarify the effect, authors investigate records of wind velocity measured for ten years (1981-90) at seven meteorological observatories. The original record is divided into ten records of velocity measured for one year. As the original one is a record of wind velocity measured at intervals of three hours, we regard each one as a sample when measurement interval is three hours. For a sample we calculate such set of parameters as mean and standard deviation which corresponds to the distribution of elements. From ten values of a parameter, authors estimate sample distribution of it when the interval is three hours. By the similar way the sample distributions of it when wind is measured for one year at intervals of 6, 9, 12, ...hours are estimated. Authors make it clear by investigation of the distribution that the scatter range of the distribution is constant in the case the interval is less than 9 hours. This means that the reliability of a parameter estimated by a result of measurement at intervals less than 9 hours is nearly equal to the value for 9 hours interval. In order to study difference of the above character between the wind velocity and the response of ships to ocean waves, records of wave induced stress measured on a pure car carrier and a bulk carrier are investigated by the same method. By the study, the same relation between the interval and reliability is also found in the case of wave stress. The only difference between them is the critical interval for the stress amounts to above 12 hours. It is concluded by the above study that the optimum interval of measurement of the response is 8-12 hours. The value may be increased by more detailed study.