Learners' ability in dealing with socio-scientific issues has been highlighted in contemporary science education. This study explored the effects of different on-line searching activities on high school students' cognitive structure outcomes and informal reasoning outcomes. By using a quasi-experimental research approach, thirty-three students were assigned to a "guided searching task group", while thirty-five students were assigned to an "unguided searching task group". The treatments of this study were two different on-line searching activities. All the participants were asked to search relevant information regarding nuclear power usage on the Internet during the period of two classes (100 min). However, the students in the un-guided searching task group were asked to search freely, while those in the guided searching task group were provided with a searching guideline. The participants' cognitive structures outcomes as well as their informal reasoning outcomes regarding nuclear power usage were assessed before and after the conduct of on-line searching tasks. The results of ANCOVA revealed that the students in the guided on-line searching task group significantly outperformed their counterparts in the extent (p less than 0.01) and the richness of their cognitive structures (p less than 0.01). Also, they significantly outperformed their counterparts in the usage of the two information processing strategies, "comparing" (p less than 0.05) and "inferring or explaining" (p less than 0.05). Moreover, it was also found that the students in the guided on-line searching task group only outperformed their counterparts in their supportive argument construction (p less than 0.05). In other words, the guided searching tasks did help the students obtain better cognitive structure outcomes; however, the increments on their cognitive structure outcomes may only help them to propose more supportive arguments, but their rebuttal construction (an important indicator for their reasoning quality) was not particularly improved.