The main achievements and innovations of this dissertation are as follows:1. The initial design of the indoor localization system does not consider different types of industrial scenarios and the performance of localization is seriously degradedunder the condition of the complex industrial environments. It is difficult to coordinate the balance between localization requirements and system resources. Therefore, on thebasis of Smart Identifier NETwork (SINET), this dissertation proposes an indoor localization framework for complex industrial environments. Through the construction of service support mechanisms, resource adaptation mechanisms and perceptron authentication mechanisms, the flexible adaptation of LBS and system resources is realized and the harsh demands of customized localization requirements for system performance are satisfied. The system framework lays the foundation for the optimization of indoor localization in the industrial environment. On the basis of system framework, a prototype system is implement in practical industrial environment and it will provide deployment and verification for the research content of subsequent chapters.2. Interfered by factors such as metal shielding, electromagnetic radiations, dust and interference from other industrial equipments, the difficulty of the synchronizationbetween nodes is increasing, which enlarges the TDOA measurement errors and reduces localization accuracy. Hence, an asynchronous localization model is proposed toimprove the accuracy of the localization by reducing measurement errors in complex industrial environment. From the perspective of parameter measurement and locationestimation, the influences of industrial environment on the accuracy of Time Difference of Arrival (TDOA) model are analyzed. Then, the asynchronous perception model andTDOA measurement algorithm are designed to reduce the influence of synchronization on parameter measurement. With the consideration of node position errors, theasynchronous location estimation algorithm is proposed to improve the accuracy of the localization model. Experiments show that the localization model proposed in thisdissertation can satisfy varied measurement accuracy requirements and improve localization accuracy.3. The frequent movement of the equipment seriously affects the link switching and data transmission of wireless nodes in the industrial production process, resulting insevere network conflicts and an extremely low successful rate of localization. Therefore,a network resource adaptation strategy for multiple-area mobile localization is proposed,which improves the location reliability by implementing efficient switch links between different areas. First, the reliability and other factors considered in constructing theindustrial indoor localization system network are analyzed. Then, the spatially complex localization scenario is divided into multiple areas, and the network model is establishedfor multiple-area mobile localization. After that, the area perceptron algorithm based on signal strength is proposed and the industrial equipment could rapidly perceive where itis located in general. On this basis, a structure of a localization superframe is designed and a resource adaptation strategy is proposed to improve the reliability of localizationdata transmission. The simulation and the testing show that the proposed network model and resource adaptation strategy can effectively improve the successful rate oflocalization in complex industrial environments and improve the reliability of localization after area expansion.4. The information of localization is closely pertinent to specific scenarios and services in the industrial process. However, the data collection is incomplete in thecomplex industrial environment. It is of difficulty for current indoor localization system to make full use of the localization information and obtain stable location coordinates.Therefore, an anomalous data processing model is designed to complement the localization data. It can increase the stability of the localization with information features such as industrial scenarios and industrial services. First of all, the device state perception model is proposed and the TDOA perception model is established on the basis of the status features and attention mechanism so as to improve the predicting accuracy of the data. Then, a location estimation algorithm based on fuzzy comprehensive evaluation method is constructed to improve the stability of location estimation with the measurement and prediction results of TDOA. Finally, the model is tested and verified in the prototype system. The results show that the data processing model proposed in this dissertation could efficiently predict the anomalous TDOA in localization data sets and improve the stability of localization.Through the research and optimization of indoor system for complex industrial environment, this dissertation provides a feasible solution for indoor positioning systems and an exploration of key technologies in industrial environments.