Abstract: Grain size is a proxy commonly used for Quaternary paleoclimate and paleoenvironment reconstruction. The Mastersizer 2000 laser analyzer is widely adopted to measure the grain size. The facility is advantaged by fast speed, high precision and good repeatability. However, the sampling introduction of it is not quantified and there may occur great randomness and uncertainty prone to human errors. In order to study the possible influences of sample amount and sample introduction methods on the test results, six types of representative Quaternary sediments were selected for repeated experiments, respectively with methods of dumping, pipette and quota sample introduction after pretreatment. The results show that the more uniform the grain size component, the smaller the difference in various sample introduction methods. If the sediment contained multiple types of grain size components and the pretreated sample amount was much enough, the grain size result of the dumping method was prone to be finer, while that of the pipette method was often coarser, and the quota sample introduction method resulted in between. Loss-of-ignition was used to calculate the content of carbonate and organic matter in the sediments. Upon the principle of the Mastersizer 2000 laser grain size analyzer and the results of repeated quantification experiments, we calculate the sample amount needed for one test. It is found that the needed amount has positive linear correlation with median grain size and impurity content. Afterwards the linear relation was calculated. Based on the universal linear relation, the median diameter range and the content of carbonate and organic matter and other impurities in the Quaternary sediments, the sample amount needed for one test could be calculated. Pouring all the pretreated samples into the test beaker could effectively avoid the unknown impacts on experimental results caused by sample amount and sample introduction methods.