We investigate the empirical properties of Chinese city airline network, where nodes and edges denote cities and direct fights. The present network is a small-world network with short average path length and high degree of clustering, and its degree distribution follows a double power law. The degree-degree correlation of the present network is different from that of the world airport network and the North American air network: The small-degree nodes of the world airport network and the North American air network are positively correlated, whereas those of the present network don't exhibit correlation behavior. And for large-degree nodes, the average neighboring degrees approach a constant in the world airport network, while the present network shows negatively correlated behavior. Previous studies suggested the absence of hierarchical structure in geographical networks, such as the Internet, the electrical power grids, and so on. However, the present network displays an obviously hierarchical property, which provides us an evidence that the geographical ingredient may have no effect on the structural evolution of Chinese city airline network. Furthermore, by setting the number of available seats per week as the edge weight, we investigate the correlation between weight and degree, and find a power-law weight-degree correlation with exponent 1.37.