In future 6G wireless communication, the utilization of ultra-massive MIMO and the requirement of high-mobility communications bring more complex propagation features to the channel. These features make it require additional overhead and higher complexity to obtain channel information using traditional channel estimation methods. To address the complexity and overhead aspects, artificial intelligence (AI), which excels at handling high-complexity problems, has attracted the attention of researchers. If we want to use AI to acquire channel information in the system efficiently and accurately, a configurable dataset with key features of 6G channel is essential. In this paper, a system parameters configurable channel dataset is proposed for future 6G research. Some key propagation features in wireless channel, such as spatial non-stationary (SnS) feature and mobility feature, are achieved in our dataset by special simulation methods and synthesis methods. Meanwhile, the simulation results are also provided to validate these features in the dataset.