The coarse alignment process is a crucial step for the Strapdown Inertial Navigation System (SINS). Achieving high accuracy in coarse alignment results can significantly enhance the accuracy of subsequent integrated navigation processes. As sensor biases play a pivotal role in the performance of the initial alignment process, addressing the elimination of sensor biases is a key research focus. In this paper, an improved alignment method for the initial alignment process, which is based on the apparent motion model of the gravity, is proposed. In contrast to prior studies, this method incorporates sensor biases into the constructed parameter model, enabling their estimation during the coarse alignment process. By employing parameter identification techniques, the accuracy of the coarse alignment process is enhanced. Simulation tests are conducted for validation, demonstrating accurate estimation of biases and effective suppression of alignment errors.