In heterogeneous cloud radio access networks (HCRAN), various types of hyper-dense cells coexist and overlap with one another. Cell selection (CS), i.e., each mobile equipment (ME) always selects the most appropriate cell for access, becomes one of the key issues that have not been well solved yet. With traditional schemes, CS decision should be triggered at any time when an ME moves across any cell border. Since the deployment of femtocells in the near future will be extremely dense, an ME may quite frequently move across the border of femtocells, causing too frequent triggers of CS decisions and resulting in additional computation, signaling, energy, and latency costs. In this paper, the priorities of various types of cells are decided by cell features, ME mobility features, and application features. Since the above features change relatively slow, the priorities can be considered as static in certain period. When moving across any cell border during such a period, instead of triggering a CS decision, our scheme is to select directly the type of cells with the highest priority. Simulations show that our scheme decreases the rate of CS decisions for MEs moving in dense HCRAN scenarios.