Abstract: Abstract: Rice harvesting mechanization in China is developing rapidly in recent years. In 2013, the total of rice and wheat combine harvester was 1.13 million, the mechanical rice harvesting area was 23.95 million hm2, of which the area cross the district was 7.70 million hm2, and at the same time, there was an area of 5.65 million hm2 harvested by manpower. Nowadays, the machine harvesting service of rice in China has been still in a supplier market, which results in the situation of making farmers wait for machines and thus prolongs the harvest time and causes a loss to the machine harvesting. However there are few researches on the loss from the machine harvest of rice and the elements of grain loss in the current studies. In this paper, the author sets a mechanical harvesting test in a paddy field with unified management during 12 consecutive days, which is to test the weight of 1 000 grains of rice, the weight of grain remained inside the harvester, the weight of grain dropped from the hit by cutting table and the weight of harvested rice, and then calculate the relative loss rate and absolute loss rate according to the data. The results show that the distribution range of the loss rate of dry matter is 0-5.88%, that of the entrainment cleaning loss rate is 0.61%-0.82%, and that of the loss rate of grain in the field is 0.21%-0.51%. The entrainment cleaning loss on different harvesting date has a significant difference at 0.05 significant level, and the loss of grain in the field and the weight of harvested rice have significant difference at 0.01 significance level. It can be seen from the F-value that the impacts of harvesting date on different factors are as follows: the weight of rice harvested > grain dropped loss in the field > entrainment and cleaning loss. On November 4 th, 5 th, 15 th and 16 th, there was no significant difference for the weight of harvested rice; on November 9 th, 10 th, 11 th and 13 th, there was no significant difference for the weight of harvested rice; on November 17 th and 18 th, there was also no significant difference for the weight of rice harvested. On November 9 th -13 th, the weight of harvested rice was the highest, and the loss rate was the lowest. The timeliness loss of rice mechanical harvesting is a quadratic curve; the longer the harvesting period, the more obvious the loss rate. The coefficient of timeliness loss of rice mechanical harvesting is 0.0009. About 5 days earlier or later from the optimal harvest date, the average loss rate of rice mechanical harvesting is 1.38%, which accounts for 30.43% of the total loss, and the average loss rate of dry matter is 0.42%; about 15 days earlier or later from the optimal harvest date, the average loss rate of rice mechanical harvesting is 3.61%, which accounts for 72.30% of the total loss, and the average loss rate of dry matter is 2.61%. In a conclusion, the optimal harvesting day is from November 9th to 13th. This research can be helpful to determine the optimal harvesting period, and produce the dynamical parameters to provide the data support for harvester allocation optimization that is aimed at the best production efficiency.