Purpose To determine the optimal cut-off score for the Modified Ashworth Scale (MAS) corresponding to unfavorable outcomes for mobility and walking ability. Methods The level of plantar flexor muscle spasticity and the 10-meter walking test (10mWT), timed up and go (TUG), and five time sit-to-stand (FTSTS) outcomes were evaluated in individuals after stroke. The correlation between MAS and the tests was investigated, and the optimal cut-off score, sensitivity, and specificity were evaluated through receiver operating characteristic (ROC) curve. Results Twenty-one participants with chronic stroke and plantar flexors spasticity (11 men; 10 women; mean age = 57.6 ± 12.5 years) participated in the study. Significant correlations between MAS and 10mWT (r= −0.45; p < 0.05), MAS and TUG (r = 0.48; p < 0.05) were found. The optimal cut-off scores were MAS > 2 for unfavorable 10mWT (sensitivity = 100%; specificity = 54.5%; ROC = 0.782) and MAS ≤ 2 for favorable TUG outcomes (sensitivity = 55.5%; specificity = 91.6%; ROC = 0.782). Conclusions This study revealed that moderate level of plantar flexors spasticity results in the highest sensitivity to predict poor gait speed performance and the highest specificity to predict good mobility performance in individuals after stroke. These findings will help clinicians in their evidence-based decision making on the role of spasticity for mobility and walking ability. Implications for rehabilitation Moderate level of spasticity (MAS <2) is the optimal cut-off score for 10mWT and TUG tests. Reducing the level of spasticity of plantar flexors below this cut-off point might be associated with an increased walking speed in this population. MAS <2 might not limit walking and mobility in individuals after stroke. Calf muscles spasticity might not compromise five time sit-to-stand (FTSTS) performances and might be related to a smaller influence on the sit to stand task.