ABSTRACT Background: In spastic paresis, the respective contributions to active function of antagonist hypoextensibility, spasticity, and impaired descending command remain unknown. Objectives: We explored correlations between ambulation speed and coefficients of shortening, spasticity and, weakness for three lower limb extensors.Methods: This retrospective study identified 140 subjects with chronic hemiparesis (>6 months since injury) assessed during a single visit with barefoot 10-meter ambulation at comfortable and fast speed, and measurements of passive range of motion (XV1), angle of catch at fast stretch (XV3) and active range of motion (XA) against the resistance of gastrocnemius, rectus femoris, and gluteus maximus. Coefficients of shortening (CSH=[XN-XV1]/XN; XN, normal expected amplitude based on anatomical values), spasticity (CSP=[XV1-XV3]/XV1), and weakness (CWK=[XV1-XA]/XV1) were derived. For each muscle, multivariable analysis explored CSH, CSP, and CWK as potential predictors of ambulation speed.Results: Ambulation speed was 0.62±0.28m/s (mean±SD, comfortable) and 0.84±0.38m/s (fast) and was correlated with CSH and CWK against gastrocnemius (CSH, comfortable, ns; fast, β=−0.20, p=.03; CWK, comfortable, β=−0.21, p=.010; fast, β=−0.21, p =.012), rectus femoris (CSH, comfortable, β=−0.41, p=6E−7; fast, β=−0.43, p=5E−7; CWK, comfortable, β=−0.36, p=5E−5; fast, β=−0.33, p=.0003) and gluteus maximus (CSH, comfortable, β=−0.19, p=.02; fast, β=−0.26, p=.002; CWK, comfortable, β=−0.26, p=.002; fast, β=−0.22, p=.010). Ambulation speed was not correlated with CSP.Conclusions: In chronic hemiparesis, ambulation speed correlates with coefficients of shortening and of weakness in lower limb extensors, but not with their spasticity level. This may encourage therapists to focus treatment primarily on muscle shortening by stretching programs and on impaired descending command by active training.