To address the surveillance problem of micro unmanned aerial vehicle ( UAV) under low signal-to-noise ratio ( SNR) environment, this article proposes an integrated target tracking and detection method based on the sequential Monte Carlo track-beforedetect (SMC-TBD) algorithm by utilizing multiple input multiple output radar. Different from conventional methods considering detection and tracking processes independently, the proposed method relies on the raw unthresholding radar data cube after 3D FFT directly to calculate the accumulative existence probability of the target. In this way, the continuous detection and high precision tracking of micro UAV are achieved simultaneously. The novelty of the proposed method is that it can realize the target energy accumulation of time-rangeDoppler-azimuth domain by integrating the detection and tracking process. Therefore, the micro UAV surveillance performance under low SNR condition is improved. Experiment results show that the micro UAV tracking performance of the proposed method deteriorates gradually only when SNR is lower than - 20 dB, which can realize 8 dB improvement compared with radar measurements, extended Kalman filter and particle filter.