Significant discrepancies among anaphylaxis severity scores may delay epinephrine administration. The study aims to develop a transparent severity grading system of foodinduced acute allergic reactions with decision model for epinephrine use. The natural course of 315 acute food-induced allergic reactions in children hospitalized at the Allergology Department between May 2016, and July 2019 due to follow-up treatment and/or allergy diagnostics, was evaluated. The severity of episodes was classified according to the five most accepted grading systems. The inter-rater reliability of classification between anaphylaxis severity scores was assessed. All symptoms were grouped into a heat map according to their real-life incidence and clinical relevance. Based on the heat map analysis, a severity grading system of foodinduced acute allergic reactions in children,with the epinephrine administration decision model, was created. Data from 259 food-induced anaphylaxis episodes in 157 children were included in the analysis. Comparing the grading systems, we observed from 24.7% to 70.2% disagreement between severity scores. The heat map showed a strong association between 29 symptoms and their categorization. A new severity grading system was developed and a two-stage decision model was proposed: "epinephrine yes" (any rapidly progressing symptoms, even mild or from one organ system; any symptoms from more than one organ system; every grade of anaphylaxis), and "epinephrine available and prepared to use" (non-progressing mild systemic allergic reaction from one system area only; no anaphylaxis). A new severity grading system of food-induced acute allergic reactions in children could serve as a clinical tool for healthcare professionals to avoid epinephrine administration delay.