Briefly, our innovation is putting a sensor on the pillow to record heart data of the users and then send the data to computer AI model to analyse it to output a potential heart disease emergency, and notice the users if they will have risks at heart later on. About its technology side or working principle, first, the basic data is measured through the sensor piece. Specifically, we aim to use BMD 101. At present, some businesses have developed BMD101 sensor modules, which basically use Bluetooth to directly transmit data to PC or mobile phone. And the positive and negative electrode sensor pieces measure heart rate by detecting the electrical signals generated by the heart as it beats. These sensors that embedded in the pillow’s neck area,establish contact with the user’s skin. When the heart pumps, it produces tiny electrical impulses that propagate through the body. The positive and negative electrodes capture these impulses, sensing the differences in electrical potential at their contact points on the skin. This method that known as bio-electric signal measurement enables the detection of the heart’s rhythm and rate. Then, the data is transmitted to the computer in the form of Bluetooth transport. The port of computer is created using python. Data is inputted into the computer and analyze by the artificial intelligent models which are trained in the way of 1D-CNN by massive original problematic heart data provided by hospital. 1D-CNN is a convolutional neural network for processing one-dimensional sequence data. It typically consists of multiple convolutional and pooling layers, and finally uses a fully connected layer to map the extracted features to the output. Finally, if AI finds some data that is common to problematic ones, computers will output the potential result and summaries, express through diagrams and description, and show on the screen of computer. If it's necessary, or too risky, it will alert the users and call the ambulance. The data is stored in such a way that the suspected abnormal beats are divided into one folder and the normal one is in another folder. According to the professor Wu, the accuracy is about 83.7%.