Introduction: The present study aimed at creating andverifying a workflow to perform subject-specific strength-scaling ofmusculoskeletal models, and validating the strength-scaled models usingisometric joint torque measurements. Methods: The participantsconsisted of 21 males and 7 females. A field strength assessment across 10exercises was used to estimate the participants’ one-repetition-maximum (1RM).The 1RM measures were implemented in an optimization routine, calculating a setof strength factors capable of scaling all included muscles in the 10 differentexercise-specific musculoskeletal models. Further, peak joint torques wereinvestigated using dynamometer obtained isometric strength measurements forelbow flexion and extension, knee flexion and extension, and ankle plantarflexion. Results: The optimization based strength-scaled models showedan improvement of mean normalized root mean square error from 48.39 (+/-22.99) % to 28.13 (+/- 15.47) % compared to the standard-scaled models.Discussion: The optimization routine was faster than previously usedmethods and showed a higher accuracy than the standard strength-scaling ofmusculoskeletal models. Issues in the simple muscle models wrapping around theknee and ankle joints made the comparison with the dynamometer data infeasible.The present study shows an improvement when applying the optimization routinefor whole body musculoskeletal models, and other or more exercises could easilybe implemented for scalability. However, utilizing simple musculoskeletal musclemodels cannot readily be used to estimate and compare peak joint torque for nearend range of motion angles.