Optimistic bias means people tend to consider that they are more likely to experience positive events and less likely to experience negative events. Pessimistic bias, in contrast, means people think that they are less likely to experience positive events but more likely to experience negative events. There are two methods to measure optimistic and pessimistic bias: direct measure (participants make direct comparisons) and indirect measure (participants make indirect comparisons). The current study explored optimistic and pessimistic bias about general life events and accidents in university students, and tested whether the measurement method would influence the results. 273 university students were invited to complete a self designed questionnaire that included 7 negative life events, 7 positive life events, 6 accidents and 6 keeping safe in accidents. 132 of them made direct comparisons, the other 141 ones made indirect comparisons. The results showed that: (1) In both direct comparisons and indirect comparisons, participants expected that negative life events and accidents were more likely to occur to others than to themselves (optimistic bias) and keeping safe in accidents were more likely to occur to others than to themselves (pessimistic bias). However, they showed optimistic bias about positive life events in direct comparisons but pessimistic bias in indirect comparisons. (2) In both two methods, participants displayed optimistic bias about accidents much stronger than about negative life events. However, there was significant difference between the bias about positive life events and keeping safe in accidents when using the direct measure, but no difference in the indirect measure. (3) In direct comparisons, participants had stronger optimistic bias about infrequent/negative events but pessimistic bias about infrequent/positive events. However, these relationships were not significant in indirect comparisons. Although the indirect comparison was proved to be more accurate in measuring optimistic and pessimistic bias, the direct comparison had been shown could predict behaviors and affective outcomes more effectively. These results suggested that the choice of measurement method should base on the research purposes.