The paper presents an efficient and reliable methodology for crystal structure prediction using a combination of *ab initio* total-energy calculations and an evolutionary algorithm. This method can predict both the most stable crystal structure and a number of low-energy metastable structures for a given compound under any pressure-temperature (P-T) conditions without requiring any experimental input. The method has achieved an extremely high success rate, with successful predictions for ionic, covalent, metallic, and molecular structures up to 20 atoms in the unit cell. The authors discuss the physical reasons behind the success of this methodology and highlight its applications in resolving important problems in high-pressure crystallography, such as the discovery of new high-pressure crystal structures for ε-oxygen, a new phase of sulphur, and new metastable phases of carbon, sulphur, nitrogen, and CaCO₃. The method is implemented in the USPEX code, which is designed to be highly efficient and scalable, capable of handling systems with up to 40 atoms in the unit cell. The paper also includes detailed tests and applications, demonstrating the method's ability to predict stable and metastable structures for various elements and compounds under different P-T conditions.The paper presents an efficient and reliable methodology for crystal structure prediction using a combination of *ab initio* total-energy calculations and an evolutionary algorithm. This method can predict both the most stable crystal structure and a number of low-energy metastable structures for a given compound under any pressure-temperature (P-T) conditions without requiring any experimental input. The method has achieved an extremely high success rate, with successful predictions for ionic, covalent, metallic, and molecular structures up to 20 atoms in the unit cell. The authors discuss the physical reasons behind the success of this methodology and highlight its applications in resolving important problems in high-pressure crystallography, such as the discovery of new high-pressure crystal structures for ε-oxygen, a new phase of sulphur, and new metastable phases of carbon, sulphur, nitrogen, and CaCO₃. The method is implemented in the USPEX code, which is designed to be highly efficient and scalable, capable of handling systems with up to 40 atoms in the unit cell. The paper also includes detailed tests and applications, demonstrating the method's ability to predict stable and metastable structures for various elements and compounds under different P-T conditions.