利用全球定位系統 (GPS) 進行測量，有許多優點是傳統測量所不及的。然而在處理長基線資料時，由於部份無模式化誤差之影響，使模稜參數 (Ambiguity parameter)成功求定的機率降低很多。因此，為了解決這個問題，吾人提出以多參考站模式化電離層影響量及增加觀測量方程式。模式化電離層方面，有兩種模式化方法被測試：其一，根據各參考站和待定站之坐標，採線性內插之方式估計二次差電離層影響量；其二，根據斜映函數 (Slant function)，採最小二乘法估計二次差電離層影響量。 另一方面，一個含有數個參考站的觀測網可提供數個已知坐標，這些參考站便與待定站組成數條基線，如此一來，自由度因基線觀測量方程式增加而增加；而這些基線將置於同一矩陣系統處理。最後以三個實驗資料作為驗證；結果顯示，不論是只使用單頻L1資料或是使用雙頻L1，L2資料，吾人所提出之概念對模稜參數求定和衛星測量都有益助。 本論文之主要目的，在於初步評估多參考站對衛星測量的效益。 Satellite surveying with the Global Positioning System (GPS) has shown several advantages over standard surveying techniques. However, because of some unmodeled systematic errors, the possibility of successful ambiguity resolution is poor in the processing of long baseline vectors. Therefore, the methods of modeling ionospheric errors and increasing measurement equations are proposed with multiple reference stations. There are two proposed method of estimating ionospheric delays; one is linear interpolation according to the positions of reference and rover stations, another is the least squares method considering a slant function. On the other hand, a network with several reference points can provide several known positions to form more baselines to rovers, so more degrees of freedom are available by increasing measurement equations of baselines. These baselines would be processed in a matrix system. Finally, for verification, three experiments were carried out. The results show that the proposal concept is helpful to ambiguity resolution and GPS surveying both with L1 frequency and dual-frequency data. The purpose of this thesis is to evaluate the help of multiple reference stations for GPS surveying.