A method is proposed for estimating two instantaneous mortality coefficients, and rate of effective survivors to number of tagged fish released, using exact times at liberty from multiple sets of tagging experiments. Effective survivors are defined here by the product of survivors immediately after releasing and reporting rate of recaptures. A least-squares method was applied to estimate the parameters using statistical models expressing the numbers of effective survivors, which were calculated by a principle similar to virtual population analysis. In order to investigate the statistical performance of the method, simulation trials using operating models were carried out. The method was applied to data of tagging experiments for alfonsino Beryx splendens from 1957 to 1969, off Kanagawa Prefecture. In most of the results of simulation trials, the method yielded estimates of natural mortality coefficient of which expected values were within ±12% of true values. Estimates of natural mortality coefficient, fishing coefficient and the rate of effective survivors for alfonsino are 0.57 (per year, 95% confidence limits [CL] 0.40-0.69), 0.094 (per year, 0.03-0.25) and 0.39 (0.12-1.00), respectively. Modifications for the generalization of the method and notes on its applications are discussed.