OBJECTIVE: Recent studies, have shown a strong relationship between maximal half-squat strength and movement velocity (Wisloff et al, 2004) and suggested that training using heavy weights (>85% 1 RM) may be preferable for soccer players (Hoff and Helgerud, 2002). However, no study has compared this type of training with a program using lower loads commonly used for resistance training in soccer. The purpose of the present study, which had Ethical Committee approval, was to examine the outcomes of two different resistance training programs (half-squat) performed 3 times/week for 6 weeks during the pre-season period. One program was designed to promote muscle hypertrophy (H, 4 sets x 12 reps, with 70% 1RM) and the other aimed to increase maximal strength (S, 4 sets x 5 reps, with 90% 1RM). METHODS: Eighteen male soccer players were divided in two equal groups. The force-velocity characteristics [maximal force at zero pedal speed (Fo) and maximal pedal speed (Vo)] of each player were determined using short maximal sprints on a Monark cycle ergometer against different loads (Arsac et al. 1996). Maximal half squat strength and fieldtest performance was measured before and after training. RESULTS: Maximal squat strength increased significantly more in the S compared to the H group (9.9±1.2% and 17.3±1.9%). Lean leg volume was increased only in the H group (by 4.3±0.8%), but was unchanged in the S group. Fo was increased only in the S group. Improvement in squat strength was correlated with improvement in 10 m sprint time (r=0.67 P<0.01) and vertical jump (r=0.63 P<0.01). DISCUSSION: The increase in maximal strength in the S group, without an increase in lean leg volume would imply that strength was increased due to neural adaptations. Expressing strength gain per unit lean leg volume resulted in a 3 to 5-fold greater increase in strength in the S compared to the H group. These results suggest that resistance training using high loads may be preferable for soccer training.