The objective of this research was to determine if tests of lifting strength and anthropometry could be used to assess training-induced changes in box-lift performance. Ninety-two British Army recruits completed a 10-week training programme. Isometric and dynamic lifting strength, body composition and maximal box-lift to 1.45 m and 1.70 m were assessed before and after training. Training significantly increased box-lift performance, dynamic lifting strength and fat-free mass, and reduced body fat. Statistically significant regression models (adjusted R2 of 0.299 and 0.152) were produced that predicted the changes in box-lift performance from changes in body composition and lifting strength. However, the 95% limits of agreement between the observed and predicted change in box-lift performance showed wide margins of error (11–13 kg; 26–29%). Training-induced changes in box-lift performance appear to be highly specific, which suggests that both training and testing procedures should also be specific to box-lift performance.