Abstract: As an important error source in ionospheric modeling and navigation, so it is very important to estimate and solve it. In order to improve the accuracy of differential code bias (DCB) estimation and ionospheric modeling, this paper proposes a multi-factor stochastic model based on altitude angle, the distance between the station and the satellite and station latitude, and analyzes the influence of different stochastic models on the accuracy of DCB estimation and ionospheric vertical total electron content (VTEC) modeling. The results indicated that different stochastic models may produce about 0.2 ns difference in satellite DCB. Compared with altitude angle weighting method, the estimation accuracy of station DCB was improved by 0.13 ns on average, and the ionospheric modeling accuracy was improved by about 0.2 TECU when the combine model of the elevation angel and the distance from the satellite to the ground station was used. Under the new stochastic model, the DCB solution of low latitude stations was worse than the first two methods, but the DCB solution of high latitude stations was better, and the modeling accuracy of ionospheric VTEC was improved significantly, which was 0.88 TECU and 0.68 TECU respectively, compared with the first two stochastic models.