With the system upgrading and construction demand for new generation global navigation satellite system (GNSS), the navigation signal modulation and multiplexing technology have made great progress. Up to now, many modulation modes for single signal component and many constant-envelope multiplexing methods for multiple signal components have been proposed, meanwhile the new signal structure continues to be presented. The satellite navigation signal code delay estimation error is the most critical factor that determines system service performance. Therefore, it is urgent to give an overall performance evaluation on code delay estimation error of GNSS signals with different modulation and multiplexing modes, and consequently provide a crucial selection basis for subsequent system application. The code delay estimation error is related to not only signal structure, but also receiver's receiving model and processing method of code tracking loop. The receiving models of new generation navigation receivers can be classified as two types. One is matched receiving model which means that the reference signal is the same as receiving signal, and the other is unmatched receiving model, where the reference signal is not the same as receiving signal. Recently, the unmatched receiving model has been extensively applied to the processing of binary offset carrier class signals. Therefore, in this paper we propose an integrated evaluation method for code delay estimation error of navigation signals. Firstly, the equivalent model for the code tracking loop of navigation receivers is generalized and the current navigation receivers are classified as four types based on whether matched receiving or coherent processing is used. Because the code delay estimation error is dependent on the type, it is necessary to provide an evaluation method for each type. Then, on the assumed condition that the code delay estimation error is very small, the expressions of code delay estimation error for coherent processing and non-coherent processing under matched receiving model are respectively presented and the relationships between each other are discussed under various noise environments and the code loop interval going to zero. The expression of code delay estimation error for non-coherent processing under unmatched receiving model is derived and the relationship with coherent processing is discussed under the same condition as matched receiving model. Finally, the Ziv-Zakai bound of code delay estimation error is derived, which provides a perfect evaluation method when the code delay estimation error is not very small. The proposed method is expressed by power spectrum density of navigation signals, which provides important theoretical guidance for signal design and receiver development, and simultaneously brings great convenience to the evaluation of the signal. Simulation experiment attests to effective evaluation on the code delay estimation error of new generation typical navigation signals.