The decoy-state protocol enables quantum key distribution (QKD) systems to achieve high performance without using the single-photon source. In the dozen years since the decoy-state protocol has been proposed, a series of advances in scheme design, finite-size analysis, system modelling, and parameter estimation has developed. Unfortunately, most advances are based on different starting points and lack a synthesis to figure out the optimal protocol of decoy-state method. Here, the advances in decoy-state method are reviewed and they are synthesized to compare the key-rate performance of 38 decoy-state protocols using the particle swarm optimization (PSO) algorithm. To form a guideline for practical QKD systems, the optimal decoy-state protocols are found with the change of five system parameters (the transmission loss, the block size of postprocessing, the misalignment-error probability, the dark-count probability, and the afterpulse rate). It is expected that this work can answer the questions of what the actual performance is for different decoy-state protocol in various scenarios and which decoy-state protocol is the optimal choice for various QKD systems.