This study developed a smartphone application that provides wireless communication, NRTIP client, and RTK processing features, and which can simplify the Network RTK-GPS system while reducing the required cost. A determination method for an error model in Network RTK measurements was proposed, considering both random and autocorrelation errors, to accurately calculate the coordinates measured by the application using state estimation filters. The performance evaluation of the developed application showed that it could perform high-precision real-time positioning, within several centimeters of error range at a frequency of 20 Hz. A Kalman Filter was applied to the coordinates measured from the application, to evaluate the appropriateness of the determination method for an error model, as proposed in this study. The results were more accurate, compared with those of the existing error model, which only considered the random error.