In sexually reproducing animals, the process of mate choice by females is often mixed with the process of male-male competition. Current models of female male choice focus mainly on how females identify the higher quality of males, but neglect the effect of male-male competition on the mate choice of females. Therefore, it remains controversial what is the relative importance of two processes in forming a social bond. We propose a new 'trial marriage' model for females' mate choice. The model assumes that females unconditionally accept any male they first encounter as their mating partner, and then conditionally switch mates to a new male of higher quality than their current partner when male-male competition occurs. This model was tested in the green weevil Hypomeces squamosus by exploring how females switched mates when males' mating interference was experimentally induced. The likelihood that females switched mates, as well as their conditional acceptance criteria of a new mate, was both raised with the intensity of males' mating interference that was manipulated in an enhanced encounter rate experiment, and in male introduction or stepwise removal experiments. These experimental findings confirm that a 'trial marriage' strategy occurs during females' mate choice. Compared with other strategies, it is more beneficial for females to choose a better mate without paying the costs of identifying males as suggested by the 'trial marriage' strategy. More importantly, using the current partner quality as the conditional acceptance threshold of new mates, females can choose better males in future encounters with potential mates. In the green weevils, males' preference for larger females and the higher possibility of the largest male winning an interference are mixed together when males' mating interference reaches a higher intensity. Therefore, the consequence of a male interference will determine which male could be chosen by a female. Under this condition, conditional acceptance of the winner becomes the most beneficial strategy of females in choosing their mates. We thus suggest that the 'trial marriage' strategy would be more efficient when males' mating interference becomes the determinant factor of females' mate choice.