Purpose: To investigate the relationship between accumulated global positioning system–accelerometer-based and heart rate–based training metrics and changes in high-intensity intermittent-running capacity during an in-season phase in professional soccer players. Methods: Eleven male professional players (mean [SD] age 27.2 [4.5] y) performed the 30-15 Intermittent Fitness Test (30-15IFT) before and after a 5-wk in-season training phase, and the final velocity (VIFT) was considered their high-intensity intermittent-running capacity. During all sessions, Edwards training impulse (Edwards TRIMP), Banister TRIMP, Z5 TRIMP, training duration, total distance covered, new body load (NBL), high-intensity running performance (distance covered above 14.4 km·h−1), and very-high-intensity running performance (distance covered above 19.8 km·h−1) were recorded. Results: The players’ VIFT showed a most likely moderate improvement (+4.3%, 90% confidence limits 3.1–5.5%, effect size 0.70, [0.51–0.89]). Accumulated NBL, Banister TRIMP, and Edwards TRIMP showed large associations (r = .51–.54) with changes in VIFT. A very large relationship was also observed between accumulated Z5 TRIMP (r = .72) with changes in VIFT. Large to nearly perfect within-individual relationships were observed between NBL and some of the other training metrics (ie, Edwards TRIMP, Banister TRIMP, training duration, and total distance) in 10 out of 11 players. Conclusions: Heart rate–based training metrics can be used to monitor high-intensity intermittent-running-capacity changes in professional soccer players. The dose–response relationship is also largely detected using accelerometer-based metrics (ie, NBL) to track changes in high-intensity intermittent-running capacity of professional soccer players.