Microwave imaging (MI) has been considered as an alternative way to X-ray mammography for breast cancer detection. This paper presents a compressive sensing based holographic microwave imaging (CS-HMI) approach for diagnosing of breast cancer. A numerical imaging system is developed to validate the proposed CS-HMI approach, which includes a realistic human breast phantom and measurement model. Small breast tumour can be detected in the reconstructed CS-HMI image via Split Bregman (SB) with using 10% measurement data. Simulation and experimental results show that CS-HMI has the ability to produce high quality image by using significantly less measurement data and operation time.