Several pre-analytical factors may influence the accurate measurements of testosterone (T) and therefore, these factors must be a significant concern. This study aimed to examine the effects of 1) time of sample collection, 2) delay to centrifugation, 3) sample matrix types, and 4) device and duration of sample storage on the T concentrations. Blood samples were collected from 34 bucks of Kacang goats. For testing the effect of collection time, 12 pairs of morning and afternoon samples were collected. For testing the effect of delayed centrifugation, 24 samples were subjected to treatments: (i) centrifuged < 1 hour after collection (control group), (ii) centrifuged 6, 12, and 24 hours after collection (test groups). For testing the different sample matrix types, 10 samples were processed as serum and plasma. For testing the effect of sample storage device and duration, 60 samples were subjected to treatments: i) frozen at -20 O C (control group), ii) stored in a cooler box, a styrofoam box, and a thermos-flask for two, four, and six days (test groups). T concentrations were measured using a validated testosterone ELISA kit. Concentrations of plasma testosterone (pT) from morning samples were significantly higher compared to afternoon samples (p<0.05). Delayed centrifugation for up to 24 h decreased significantly on pT concentrations (p<0.05). The concentrations of T from serum and plasma did not differ and showed a strong correlation (r=0.981). Storage device and duration affected the T concentrations compared to frozen samples (p<0.05) which T concentrations were stable for up to 4 days in a styrofoam box and a thermos-flask and up to 6 days in a cooler box. In conclusion, the measurement accuracy and stability of T concentrations in goats are affected by collection time, delay to centrifugation, and device and duration of storage.