Abstract Background and Purpose To improve existing clinical assessments for freezing of gait (FOG) severity, a new clinician‐rated tool which integrates the varied types of freezing (FOG Severity Tool‐Revised) was developed. This cross‐sectional study investigated its validity and reliability. Methods People with Parkinson's disease who were able to independently ambulate eight‐metres and understand study instructions were consecutively recruited from outpatient clinics of a tertiary hospital. Those with co‐morbidities severely affecting gait were excluded. Participants were assessed with the FOG Severity Tool‐Revised, three functional performance tests, the FOG Questionnaire, and outcomes measuring anxiety, cognition, and disability. The FOG Severity Tool‐Revised was repeated for test‐retest reliability. Exploratory factor analysis and Cronbach's alpha were computed for structural validity and internal consistency. Reliability and measurement error were estimated with ICC (two‐way, random), standard error of measurement, and smallest detectable change (SDC 95 ). Criterion‐related and construct validity were calculated with Spearman's correlations. Results Thirty‐nine participants were enrolled [79.5% ( n = 31) male; Median (IQR): age–73.0 (9.0) years; disease duration–4.0 (5.8) years], with fifteen (38.5%) who reported no medication state change contributing a second assessment for reliability estimation. The FOG Severity Tool‐Revised demonstrated sufficient structural validity and internal consistency ( α = 0.89–0.93), and adequate criterion‐related validity compared to the FOG Questionnaire ( ρ = 0.73, 95% CI 0.54–0.85). Test‐retest reliability (ICC = 0.96, 95%CI 0.86–0.99) and random measurement error (%SDC 95 = 10.4%) was acceptable in this limited sample. Discussion and Conclusions The FOG Severity Tool‐Revised appeared valid in this initial sample of people with Parkinson's. While its psychometric properties remain to be confirmed in a larger sample, it may be considered for use in the clinical setting.