In this study, a scale regarding preservice teachers’ self-efficacy beliefs in educational internet use is developed. Theeducational internet use self-efficacy beliefs scale includes 28 survey items. Findings from the current study show that theCronbach alpha reliability coefficient of the scale is found 0.96. In the literature, the suggested level for the Cronbach alphareliability coefficient is 0.70. In the light of this fact, it can be concluded that the Cronbach alpha reliability coefficient of the scaleis high. The participants of this research study are preservice teachers during the academic year of 2008-2009 in the AhmetKeleşoğlu College of Education at Selçuk University. To develop the survey, three different study groups are formed. The firstgroup formed for the construct validity includes 367 college students. Of those participants, 44.1% (162) is female and 55.9%(205) male. In the second phase of the study, the data are collected for the criterion-based validity. In this phase, the study groupis made up of 326 students, 44.7% (146) female and 55.3% (180) male. The last study group is designed for the test-retestreliability. It includes 84 students. Among those preservice teachers, 42.9% (36) is female and 57.1% (48) male.In the literature, it is stated that students frequently use the Internet for emotional, social, and leisure time, and activities, notfor academic or area-specific reasons (Young, 1998). It is also suggested that healthy and effective Internet use is related topsychological maturity and self-efficacy (Wang, 2001). In the present study, self-efficacy beliefs are used to elaborate and extendBandura’s (1997) general social cognitive theory within the context of educational technology, specifically educational Internetuse. This theory consists of two important constructs: self-efficacy beliefs and outcome expectations. Self-efficacy beliefs refer toan individual’s performance capabilities (Lent, Brown, & Hackett, 1994) while outcome expectations refer to the expectedconsequences of an action (Bandura, 1997). In the social cognitive theory, self-efficacy beliefs and outcome expectations areconceptually related. Indeed, they are different constructs. The question of “can I do this?” usually refers to self-efficacy beliefswhile the outcome expectation might be determined through the question of “if I do this, what will happen?” It is important tonote, “as people develop an affinity for an activity at which they feel efficacious and expect positive outcomes, they form goals forsustaining or increasing their involvement in that activity” (Lent et al., 1994, p. 264).In the current study, self-efficacy beliefs serve as the theoretical framework in the process of the survey development. In therelated literature, this construct is titled “self-efficacy” (Akkoyunlu, Orhan & Umay, 2005; Deryakulu, Buyukozturk, Karadeniz, &Olkun, 2009), “self-efficacy beliefs” (Akbulut, 2006; Akgün, 2008; Aşkar & Umay, 2001; Orhan & Akkoyunlu, 2003; Köseoğlu,Yılmaz, Gerçek & Soran, 2007) or “self-efficacy perceptions” (Akkoyunlu & Kurbanoğlu, 2003). In this study, the term “selfefficacybeliefs” is preferred to determine preservice teachers’ capability in Internet use for educational purposes. Self-efficacybeliefs are an important factor to find out individuals’ computer use frequency and achievement (Cassidy & Eachus, 2002). Therelated literature suggests that self-efficacy beliefs positively affect computer use (Compeau & Higgins, 1995; Compeau et al.,1999; Hill, Smith & Mann, 1987).In the beginning of the study, a pool of survey items is formed to evaluate the university students’ self-efficacy beliefs ineducational Internet use. For this purpose, several open-ended questions such as “how can the Internet be used better ineducation?” and “in which activities should students use the Internet?” are asked to 110 students from the Department ofComputer and Instructional Technologies. Then, students’ responses are analyzed and ordered from the most repeated to theleast. At this stage, the item pool includes 37 attitude sentences. The items are evaluated with the options of “totally measuring”,