The present paper derived a method for estimating fishing mortality coefficient from the mean recovery time of tagged members within a finite period with a few closed seasons and showed anunnerical example. Using recovery data over 2 or more seasons, the derived method can give the maximum likelihood estimates of fishing and apparent total mortality coefficients regardless of the survival rate up to the beginning of the period and of the rates of discovery and reporting. The numerical example also demonstrated that the estimates thus determined gave the estimates of several other parameters on some assumptions. When the coefficients are practically estimated with the method, however, there arlses some difficulty in determining time T_n at the end of a tagging experiment. Moreover, this paper discussed the estimate of apparent natural mortality coe-fficient determined in the numerical example of 1-year-old red sea bream because it is very large.