<p>Low-Earth Orbit (LEO) based signal of opportunity positioning offers an attractive alternative for backup or augmentation of Global Navigation Satellite Systems (GNSS) due to its characteristics such as high signal power and high orbital velocity. Since the signals broadcast by LEO do not contain the signal transmission time, accurate signal propagation time estimation is crucial to improving the performance of LEO Doppler positioning systems. This paper proposes a novel outer-loop Doppler positioning framework that accounts for signal propagation time. It introduces a purely geometric model to estimate signal propagation time and incorporates the effects of signal propagation time during the construction of the observation model. The experimental results show that the proposed algorithm can effectively improve the accuracy of state estimation, and the accuracy of three-dimensional position and velocity estimation is improved by more than 15% and 25% respectively, and the accuracy of error estimation related to the receiver clock is also significantly improved. The theoretical limit of the initial state to which the model can converge is also explored. When the initial position error reaches 200&#xa0;km, the tolerable clock error drops to less than 50&#xa0;ms. In addition, experiments are carried out using measured Iridium observation. Compared with the existing algorithm that considers the propagation time error as the coarse-time parameter estimation, the point positioning errors of the proposed algorithm is reduced 13.4% and the stability is significantly improved.</p>