Abstract: The impact of ambiguity resolution on the accuracy of global navigation satellite systems (GNSS) is significant. To optimize the ambiguity resolution in real-time kinematic (RTK), a data and model-driven GNSS ambiguity resolution method is proposed. This algorithm focuses on selecting the optimal subset of ambiguities to increase the success rate of ambiguity fixing. First,considering the influence of ambiguity subsets on the accuracy of baseline solutions, selecting a subset method based on the ambiguity level to ensure a high success rate of ambiguity fixing and precise baseline solutions. Model-driven and data-driven approaches are combined when performing partial ambiguity resolution, introducing long short-term memory neural networks (LSTM) to ensure the stability and reliability of ambiguity fixing solutions. Results show that, while ensuring a high reliability of ambiguity fixing, compared to existing algorithms, this algorithm LSTM algorithm based on solving rules improves the fix rate and fix success rate of the selected dataset from 73.59% to 99.23% and from 72.83% to 95.52%, respectively, and from 65.27% to 99.69% and from 64.91% to 95.85%, respectively, this leads to enhanced accuracy and robustness of RTK positioning.