Abstract—Self-feeding robot is a smart assistive device for Parkinson’s patients and people who cannot feed themselves. The self-feeding robot includes 3 main parts: rotate tray, 2 Degree of Freedom (DoF) grip arm, remote control. This device configuration not only provides an efficient and easiest way to manufacture but also spreading out more devices with affordable price. Besides, to help patients on their activities and improve metabolism better with nutritional diets, eating recommendation with the supporting of machine learning could enhance the ability of analyzed patient’s nutritional consumption and give out diet recommendation directly instead of doctor advisor. The goal of this research was to employ a portable device to make disabled people who cannot eat by themselves now could have previously been served at meals completely or partially independent. With an independent manipulator, a rotate food tray easily control by only 2 simple buttons, highly integrated smart recommendation based on medical index and lightweight plastic fiber, self-feeding robot namely Feedbot is full of hopefully presented for handicapped person eating individuals not only particularly for Parkinson patients but also elderly and disable who have trouble in feeding their meals. Feedbot operated autonomously and wirelessly transferred meal records and relevant information to the software controlling in a mobile device. Furthermore, the reporting system can analyze data and suggest daily meals based on nutritional standards categories. Finally, Feedbot is a user-centered design product which allows to controlling robot’s grip and food tray rotates followed the need of user and the user’s height based on human anthropology.