Abstract This study aims to determine the effects of small-sided training games 3 versus 3 and small sided games 6 versus 6 on the speed and agility of soccer players. The study design was randomized experiment control group pretest-posttest design. The study population was students Bima Amora Soccer School in Gresik city age group 11-12 years amounted to 33, the population is randomly divided into 2 groups: experimental group A were given training small sided games 3 versus 3, experimental group B control group with conventional training. Each group was put through a test of speed and agility tests before and after treatment with a test run of 40 meters. Techniques of analysis in this study using t test. T test obtained in group A obtained results for speed T count = 17,175 and T table = 2.