The Artificial Internet of Things (AIoT) is considered to reshape future business models and provide accurate services for multi-source, heterogeneous, massive, low-value density IoT data, which is vital to deepening the application of IoT. Aiming at the existing data collection mechanism’s low accuracy, limited latency performance, underutilization of edge computing resources, and neglect of user data collection context, a content and context integrated data acquisition method is proposed. First, an edge entity observation content prediction method is proposed to precisely assess the entity observation based on the edge computing. Second, a cloud context-aware approach is designed that considers the user’s explicit and implicit data acquisition context to select appropriate entity data according to user preferences. Finally, an intelligent data collection mechanism that integrates edge resources and cloud resources is presented to improve the performance of AIoT data collection services. Simulation results show that the proposed method can enhance 7% of the precision and lower 16% of the delay performance in comparison with traditional methods.