This study aims to analyze the reliability of the test of 8 throws of 3 different variants of handball throws in university students. The participants were 33 males, 20 ± 1.9 years; weight 72.39 ± 8.9 kg; height 176.75 ± 6.07 cm; BMI 23.14 ± 2.3. They were evaluated in 2 sessions separated from each other by a period of one week. The participants executed throws until 8 measurements were recorded in each of the tests, unspecific throws, in support and in the race with 3 steps; the launch speed was recorded with a radar behind the goal. The reliability of the average of the test scores of 8 measurements between the first session and the second was measured with a test-retest test. The following results were obtained: in the nonspecific launch, an ICC= (0.86) was obtained between 95% of confidence, one of CV= (7.43%) and one ES = (0.08). For the toss with support, an ICC = (0.94) was obtained at 95% confidence, one of CV = (4.71%) and an ES = (-0.09). For the running throw, an ICC = (0.96) was obtained at 95% confidence, one of CV = (4.00%) and one ES = (0.05). This study has shown that the test of 8 throws and the 3 variants is a reliable and stable performance measure for university-level students without experience in handball; in relation to the throw variables, all 3 were reliable. Only the nonspecific variable was lower compared to the supported and running pitch. © 2022 Kassel University Press GmbH. All rights reserved.