When star sensors operate under near-Earth daytime conditions, the intense background radiation from the sky severely interferes with the energy of the star points in the imagery, resulting in a low signal-to-noise ratio (SNR) for the star points. This low SNR hinders target extraction and centroid positioning, thereby affecting the normal attitude measurement of star sensors. Addressing the challenge of attitude measurement under daytime conditions, this study first analyzes the mapping relationship of pixel positions in consecutive frames of star sensor imagery. A star image superposition algorithm based on attitude-related frames is proposed. On this foundation, an attitude measurement method based on star image superposition is employed for measuring the attitude of daytime star sensors. Furthermore, a fitting algorithm for the solar centroid is introduced, and a coarse measurement method based on solar position is applied to determine the optical axis orientation of daytime star sensors, enhancing their robustness in daylight conditions. The algorithm proposed in this study is validated through experiments. The results demonstrate that the multimodal attitude measurement method not only effectively improves the SNR of near-Earth daytime star sensor imagery through star image superposition, ensuring the accuracy of attitude measurements, but also ensures the robustness of attitude measurements through solar centroid fitting.