Objective To develop a safe feeding assessment scale for stroke patients with dysphagia. Methods An item pool was deve-loped and then two rounds of consultation with 15 experts, pilot survey with 30 patients, and formal survey with 151 patients were conducted. The response rate, authority and Kendall′s coefficients of concordance (W) of experts, as well as reliability and validity of the scale were calculated. Results The final scale consisted of 24 items assessing 6 dimensions, which explained 71.765% of the total variance. The I-CVI ranged from 0.867 to 1.000 and the S-CVI was 0.833.The test-retest reliability was 0.754 and the Cronbach′s α coefficient was 0.931. The response rate was 100%, the expert authority coefficient (Cr) was 0.936, and the Kendall′s coefficient of concordance (W) were 0.288 and 0.375 respectively (P＜0.01 for both). Conclusion The safe feeding assessment scale for stroke patients with dysphagia has good reliability and validity and can be used as an assessment tool.