Abstract: To investigate the precursor characteristics of fractured rocks under different stress conditions, cracks were prefabricated in cubic sandstone specimens, and two forms of parallel cracks and collinear cracks were made to represent different rock structural planes. Biaxial compression tests were conducted on prefabricated fractured sandstone under lateral pressures of 0 MPa, 5 MPa, and 10 MPa, respectively. Acoustic emission technology and digital imaging technology were used to monitor the development of internal fractures and surface deformation of the rock. By using the acoustic emission event rate function and the maximum principal strain field variation of sandstone, the development and failure process of rock fractures were studied, and the results showed that sandstone exhibited a progressive failure characteristic. The failure mode is mostly dominated by splitting failure, accompanied by compression shear failure, and the failure mode often presents a skewed "I" shape in the direction of the maximum principal stress, occasionally showing an "X" shape. The confining pressure has a "reinforcing" effect on the strength of the rock, and as the confining pressure increases, the "reinforcing" effect becomes more obvious, but at the same time, the stepwise trend of progressive rock failure decreases. At the same time, there is a strong correlation between the acoustic emission event rate function and the variation of the principal strain field on the rock surface, and their joint analysis can effectively characterize the failure process of rocks. In the compaction stage, the acoustic emission event rate is at an extremely low level. During the elastic and plastic stages, the acoustic emission event rate gradually increases and remains at a high level. The main strain field shows a trend of localization. In the failure stage, a localized band of the main strain field is formed, and the acoustic emission event rate gradually fluctuates violently. Along with the rapid development of rock fractures, the acoustic emission event rate suddenly increases. In order to characterize the precursor information of the sample, the coefficient of variation of the longitudinal strain field and the standard deviation of the longitudinal strain field were introduced to describe the precursor information of the digital image before rock failure, and the acoustic emission precursor information before rock failure was described using acoustic emission ring count. By using the entropy weight method, the weights of the acoustic emission precursor index and the digital image strain precursor index were determined, and a fusion index AD that can reflect the internal and surface information of the rock was obtained. The results show that the fusion index can well describe the precursor characteristics of rock failure. The fusion index has similar precursor warning time and acoustic emission warning time, and is earlier than the longitudinal displacement field variation standard deviation and coefficient of variation. In terms of the presentation characteristics of precursor characteristics, it is more obvious than the longitudinal displacement field variation coefficient, longitudinal displacement field variation standard deviation, and acoustic emission ringing count.