<p>Real-time sensing of precipitable water vapor (PWV) using GNSS is critical for atmospheric monitoring. Conventional real-time precise point positioning (PPP) reliant on the IGS real-time service (RTS) remains limited in oceanic regions due to terrestrial internet dependencies. BeiDou’s satellite-based PPP-B2b service overcomes this limitation, enabling real-time PWV sensing over the ocean. However, its relatively lower accuracy has hindered broader application. Additionally, the potential of combining PPP-B2b with low-cost GNSS devices for oceanic PWV sensing remains insufficiently explored. This study presents a practical and cost-effective approach for real-time oceanic PWV sensing by jointly using PPP-B2b and low-cost GNSS devices. To address the inherent errors in PPP-B2b corrections, we develop a refined PPP strategy based on ambiguity reparameterization and optimized stochastic modeling. Unlike conventional PPP that treats ambiguities as constants, this method models them as time-varying random-walk processes within an optimized range of process-noise values. This allows for adaptive absorption of PPP-B2b correction errors, thereby improving the accuracy of tropospheric estimates. Validation through land-based static and kinematic experiments demonstrates accuracy improvements of 34 and 24%, respectively, in tropospheric estimates. When applied to IGS-RTS corrections, the refined method achieves a 10% accuracy improvement in real-time tropospheric estimation. Two oceanic campaigns further validate its operational feasibility, yielding a PWV accuracy improvement of over 30% during Cruise 1 and 47% during Cruise 2 compared to conventional PPP. This framework offers a seamless, accurate and cost-effective solution for real-time tropospheric and PWV sensing, benefiting weather forecasting and climate monitoring.</p>