The aim of this research is to find a significant difference between teaching vocabulary using TPR method and teaching vocabulary using conventional method and prove the influences of both. The research was conducted in the form of quasi experiment. The researcher did the experiment using these two classes as a comparison. To analyze the data the writer used statistical calculation of the t-test to determine the final calculation of to (t- observation) that was done to measure the last score of the research test. In analyzing the data, the writer uses the comparative and the influences techniques. This techniques is to prove statistically whether any significant differences and influences between two variables. The processed and analyzed data through the following steps: describing the result of the test of each class, determining interval, mean, modus, median, deviation standard, variance of each class and the last is proving of t-test. The result of the statistic calculation indicated that t-observation (t0) is higher than t-table (tt), 5.82>2.462, so the alternative hypothesis (Ha) is accepted and the null hypothesis (Ho) is rejected. And based on the comparison result, it was found that zh = 4.64, Z0.05 = 1.645, Z0.01= 2.462, it concludes that 2.462 < 4.64 > 1.645, So Ho rejected if t0 > tα/2. According to the explanation above, we can interpret that teaching vocabulary using Total Physical Response is quite success. It can be proved by the data of students' score of experiment class is higher than controlled class.