The success probability of the Grover quantum search algorithm decreases quickly when the fraction of target items exceeds 1/4, where the phase plays a significant role. Therefore, we use multiple phases to complement each other. We obtain three useful properties and an important theorem of the success probability and design a systematic solution of the optimal phases for an arbitrary number of phases. Based on these results, we finally propose a multi-phase quantum search algorithm whose success probability rises with the increase of the number of phases with just a single iteration, and it tends to be 100% when the fraction of target items is over a lower limit.