This paper describes a comparative study of the influence of three sample preparation techniques (M1: hydration+oscillating extraction+partial extraction solution hexane partitioning cleanup; M2: hydration+oscillating extraction+overall extraction solution SPE cleanup; and M3: pure acetonitrile homogeneous extraction+overall extraction SPE cleanup) on the determination efficiency of 456 pesticide multiresidues in tea. First, it was discovered from the mathematical correlation equation of 329 pesticide recoveries established and log Kow values that the extraction efficiency of hydration method M1 has obvious correlation with pesticide log Kow, making the extraction efficiency of M1 take the shape of an arc trend line with a certain arc hanging down from both ends of polar pesticides and nonpolar pesticides. Second, regarding the M1 method, the interfering matter after co-extraction increased in large quantities, which markedly lowered the S/N of the target pesticides and method sensitivity, leading to an obvious decrease of the method efficiency. The fortification experiment of the uniform limit 0.010 mg/kg proved that with the M1 hydration method there are 23 pesticides with recoveries between 70 and 120% and RSD<20%, accounting for only 5.0%, while with nonhydration method M3 there are 229 pesticides, making up 50%.