The aim of this study was to assess the influence of the load order used (i. e. incremental, decremental or random loads order) during the bench press throw load-velocity profile on peak velocity achieved against four different loads (20–40–60–80% of one repetition maximum [1RM]). Both intraclass correlation coefficient (ICC) and coefficient of variation (CV) were calculated to assess the reliability of the measures. A repeated measures ANOVA was used to assess differences between protocols. A linear regression analysis was performed to assess the load-velocity relationships among the different protocols. Peak velocity showed good to high ICC values independently of the load used (ICC=0.83–0.92). CV scores showed good reliability (ranging between 2.2 and 6.2%). No significant differences in peak velocity attained at each load were found between the three testing protocols (p>0.05). In addition, peak velocity at each load was very large to almost perfect, correlated between protocols (r=0.790–0.920). The linear regression model showed a significant relationship between testing protocols (p<0.001; R2=0.94). In conclusion, due to some ICC scores below 0.9 and R2 below 0.95, the indistinct use of different load-order protocols to assess load-velocity relationships in the bench press throw exercise is not recommended.