OBJECTIVE: We aimed to develop the health behaviour scale for gestational diabetes mellitus patients (HB-GDM) and evaluate its psychometric properties. The scale may provide theoretical basis and evidence for identifying the pathways to enhance health behaviours and optimise health management strategies. METHODS: The initial constructs and items of the scale were developed through literature review, qualitative analysis and Delphi expert consultation based on the socio-ecological model. Item analysis was conducted by four methods and using a sample (n = 235) recruited in China to form formal scale. Additional participants (n = 505) completed survey to measure the internal consistency reliability, test-retest reliability, content validity, construct validity and criterion-related validity of scale. RESULTS: The scale contains six dimensions with a total of 29 items. The Cronbach's α of the scale was 0.912, McDonald's ω of the scale was 0.936, test-retest reliability was 0.957, content validity was 0.935, The cumulative variance contribution rate of the six common factors was 77.488 % and CFA model had acceptable goodness-of-fit indices(χ CONCLUSION: HB-GDM scale showed satisfactory psychometric properties and has strong specialisation and better applicability. PRACTICE IMPLICATIONS: The scale is favourable for pregnant women with GDM to clarify their unhealth behaviours. It provided guidance regarding maternal and postpartum follow-up care and the formulation of women health care strategies with regional characteristics.