Optical methods such as absorptiometry, fluorescence, and surface plasmon resonance have long been explored for sensing glucose. However, these schemes have not had the clinical success of electrochemical methods for point-of-care testing because of the limited performance of optical sensors and the bulky instruments they require. Here, we show that an ultrasensitive optical transducer can be used for wireless glucose monitoring via a smartphone. The optical transducer combines oxygen sensitive polymer dots (Pdots) with glucose oxidase that sensitively detect glucose when oxygen is consumed in the glucose oxidation reaction. By judicious design of the Pdots with ultralong phosphorescence lifetime, the transducer exhibited a significantly enhanced sensitivity by 1 order of magnitude as compared to the one in a previous study. As a result, the optical images of subcutaneous glucose level obtained with the smartphone camera could be utilized to clearly distinguish between euglycemia and hyperglycemia. We further developed an image processing algorithm and a software application that was installed on a smartphone. Real-time dynamic glucose monitoring in live mice was demonstrated with the smartphone and the implanted Pdot transducer.