This study presents a wireless non-contact atrial fibrillation (AFib) detection method based on a 24-GHz microwave Doppler sensor. Microwave Doppler sensors use the Doppler effect to measure the vibration velocity of an object in a non-contact manner. Heartbeat components can be extracted from Doppler signals, allowing the measured body surface vibrations to be correlated to a seismocardiogram (SCG). Our study applies this technique to the detection of AFib. Using a microwave Doppler sensor, we measured body surface vibrations of healthy subjects and AFib patients, and compared the differences. The differences in feature values obtained through frequency analysis and machine learning indicated the possibility of non-contact detection of AFib using a microwave Doppler sensor.