The dietary health of children is always a major issue in the field of health care, and one of the important parts is children's dietary assessment. However, the traditional approaches of dietary assessment are inefficient and hard to be popularized, and the data of dietary assessment are not utilized well. So the system is commissioned by Chinese Capital Institute of Pediatrics, which can solve the challenges of enhancing the efficiency of children's dietary assessment, improving the accuracy of estimation, and making full use of the data of dietary assessment. The system implements a complete process of Chinese children's dietary assessment. In this system, an approach of food weight estimation based on pictures taken by smartphones is proposed to make dietary assessment easily, an approach of error elimination based on machine learning is proposed to improve the accuracy of estimation, an approach of recommending foods to children that they need is proposed to utilize the data of dietary assessment meaningfully. The system can help hospitals and parents to assess children's diet easily, and provide them with a powerful reference to improve children's dietary health.