Objective This paper aims to evaluate data reliability of the neurotoxic hazard assessment of glutamate and its salts and provide recommendations, as well as to improve toxicology data reliability assessment method (TRAM) via trial application. Methods TRAM was used to evaluate the reliability of 60 articles which were selected by the method of systematic review documentation retrieval. The evaluation was based on types of toxicological data involved in each paper (laboratory animal data or human data) and they were scored by reliability criteria. The quality percentage was obtained via calculations to judge reliability categories and give recommendations. It’s necessary to note that the evaluation of each paper was independently completed by two persons in related fields. Results After three rounds of evaluation, the reliability of 12 articles were evaluated as “high” and recommended for priority use. The reliability of 43 articles was rated as “moderate” and can be used. The reliability of 5 articles was evaluated as “low” and not recommended to use. Conclusion TRAM takes both reporting quality and methodological quality into consideration, especially including human data reliability evaluation method which is absent in the other toxicology data reliability assessment tools. TRAM is more suitable for food safety risk assessment. It provides a better objective and scientific guarantee for hazard identification and risk assessment.