With the rapid growth of the construction industry, precise and efficient monitoring of building deformation has become a crucial factor in ensuring their secure construction and stable operation. The frequent occurrence of occlusion and diffraction errors in satellite signals caused by high-rise buildings has emerged as a primary challenge that constrains the accuracy of positioning in deformation monitoring by BeiDou Navigation Satellite (BDS). To solve the problems, this study proposes an elevation mask angle modeling method of satellite signal occlusion based on azimuth rounding and applies it to real-time BDS building deformation monitoring and positioning algorithm. The proposed approach involves defining a particular subset of integer azimuth and elevation angles, followed by utilizing the minimum elevation angle from this subset to develop a model for the azimuth-dependent elevation mask angle. We aimed to develop an elevation angle mask model using a conventional approach, and the model can be utilized to focus on non-line-of-sight (NLOS) and multipath signals and attenuate their impact on positioning accuracy by reducing their weight. The experimental results show that the elevation angle mask model can be easily completed by using the observation data of BDS satellites in a single day, and its use can make the number of effective satellites involved in the calculation of real-time deformation monitoring of BDS more accurate and reliable. The root mean square (RMS) of positioning results including float solutions was increased respectively by 91.1%, 93.3%, and 72.7% in the direction of East-North-Up (ENU), and the RMS of results with fixed ambiguity was increased by 6.9%, 10.0%, and 29.5%, respectively. Therefore, it can be concluded that the elevation mask angle modeling algorithm is convenient to realize and can significantly improve the performance of deformation monitoring results by BDS.