Consuming additional sugar can exacerbate the development of obesity and diabetes, two conditions that are becoming more and more dangerous to the general public's health. While knowledge of nutritional content might significantly impact consumption patterns, not all nations now mandate the disclosure of added sugar. Nonetheless, an increasing number of people worldwide have access to portable devices, which makes it possible to develop technological tools to support actions and decisions linked to health. This article gathered detailed nutritional data, including added sugar content data, for 69,769 foods in order to investigate whether developments in computational science can be used to create a scalable and precise model to forecast the added sugar content of foods based on their nutrient profile. A gradient boosted tree model estimating added sugar concentration was trained using 80% of the data, with the remaining 20% kept out to evaluate the model's predictive power. The final model's performance revealed that 94.89% of the variation was explained for each default portion size. For each preset portion size, the mean absolute error of the estimate was 0.90 g. Because of this, this approach can be used to provide precise added sugar estimates via digital devices in nations where the data is not given on packaged foods, allowing customers to be informed of the added sugar amount of a wide range of meals.