Japan's Quasi-Zenith Satellite System (QZSS) transmits signals at the same frequency as that of the Global Positioning System (GPS). For users in Japan, QZSS can be received for more than 12 hours a day with a satellite elevation angle of above 70°. In order to integrate QZSS with GPS to achieve precise positioning, a real-time kinematics (RTK) algorithm is proposed in this work. For the combined use of QZSS and GPS measurements, the conventional RTK algorithm works well for the short baseline scenario (i.e., baseline is less than 10 km). However, for medium-range baseline RTK (i.e., baseline is longer than 10 km but less than 50 km), systematic errors such as the ionospheric delay error grow with distance; in other words, these errors are less spatially correlated compared to that of the short baseline RTK. In order to extend the baseline length between the RTK rover and reference receiver, an improved ionospheric correction technique is proposed in this work for the combined use of QZSS and GPS. In order to enhance the availability of RTK fixed solutions, this research also proposes a satellite selection method because there are still some float solutions when the proposed ionospheric correction technique is applied. This satellite selection method analyzes all the combinations of QZSS and GPS satellites and finds the best solution for the rover station. There are many RTK system architectures. The standard RTK architecture is used in this work. It simply uses one reference station with a dual-frequency Global Navigation Satellite System (GNSS) receiver to generate corrections for the rover receiver. The main RTK algorithm combines the double difference (DD) method with the Lambda algorithm. The satellite selection method can be divided into three parts: 1) selection using the DD measurement weighting covariance; 2) selection using the DD ionospheric difference; 3) selection using the satellite elevation angle; 4) selection using the C/N0 value. Two GNSS experiment data sets are studied in this paper. The baseline lengths between the reference station and the rover are 18 and 27 km. The data sets are used to examine the feasibility of the proposed techniques. The differences of the dual-frequency code- and carrier-phase measurements between GPS and QZSS are first analyzed. An improved ionospheric correction technique is then proposed for medium-range baseline RTK that integrates GPS and QZSS. Finally, the results of the satellite selection method are discussed in detail. The results demonstrate that RTK positioning is improved by the use of the proposed method. Importantly, with the improved ionospheric correction technique and the proposed satellite selection method, the availability of the successful RTK fixed solutions for the combined systems is improved by more than two-fold (from 47% to 99%).