In this paper, we propose a systemic architecture of network selection based on context-awareness services, which gathers contextual information that includes such network information, user information and local information. This network selection strategy considers the Quality of Service (QoS) and user preferences. Also, it perceives contexts such as speed, coverage percentage and location, etc., and it eventually performs network selection decision making and network execution based on multiple factors. From the perspective of network decision, it presents two network selection algorithms, namely the fuzzy mathematics evaluation method and multiple attribute decision making using the TOPSIS evaluation method. System simulations suggest that network selection based on the mathematics evaluation method is much faster than the TOPSIS evaluation method. However, the TOPSIS evaluation method is practically more efficient. The network selection method based on context-awareness provides an effective and flexible network vertical handover strategy, and ensures a good accuracy and efficiency.