Abundance and structure of feeding aggregations of chum salmon are analyzed on the results of the otoliths processing sampled in different areas of the Okhotsk Sea in 2011-2014 and scheme of their autumn migrations in dependence on the sea surface temperature (SST) patterns, using the data on catching the fish with the otoliths marked in hatcheries. The migrations were similar for all chum groupings from the Okhotsk Sea and adjacent waters: their juveniles moved to Kamchatka coast until SST decreasing to 11оС in October, where they formed mixed aggregations, then density of the aggregations increased until SST 9-10оС, after that the aggregations moved to the central part of the sea and further southward passing the Kuril Straits in November under SST 7-8оС. The portion of marked fish from hatcheries in mixed catches was 2.1-7.8 % (on average 3.9 %). Taking into account these values and the ratio between marked and non-marked fish for certain hatcheries, the ratio between wild stocks and hatchery release of different origin is determined. The portion of chum salmon from Japanese hatcheries is estimated as 9.6-41.6 % (on average 27.3 %), the portion of chum from Russian hatcheries is considerably lower - 0.5-2.6 % (on average 1.2 %), and chum salmons of natural origin dominate in all years with the portion 55.8-89.9 %. For the fish of artificial origin, the ratio between the number of juveniles feeding in the sea and their release from hatcheries is estimated as 3.0-4.2 % (on average 3.6 %) for Japanese hatcheries and 0.3-0.7 % (on average 0.5 %) for Russian hatcheries. So, survival of chum from Japanese hatcheries is higher in 8 times on average, and total abundance of the feeding artificial chum salmons of Japanese origin is higher than those of Russian origin in more than 20 times. However, the hatcheries in Sakhalin and Kuril Islands has similar efficiency as Japanese ones, in opposite to other areas. The abnormally low estimations of the efficiency for some Russian hatcheries allow us to suppose that the otolith marking program is not realized correctly in some regions of Russian Far East, possibly the marks don’t correspond exactly to the schemes of marking.