This study’s main aim was to determine physical education (PE) teacher candidates’ self-efficacy levels and attitudes toward the PE teaching profession. Designed on a survey model, this study was conducted during the 2011-2012 academic year. A total of 601 PE teacher candidates studying in the PE and sports teaching programs of six different universities in Turkey participated in the study. Study data were collected using the “Teacher Self-Efficacy Scale (TSES)” and the “Attitude Scale for the Profession of Physical Education Teaching (ASPPET).” Descriptive analysis, t-test, Pearson Correlation Coefficient Technique, and Multiple Regression Analysis methods were used for data analyses. PE teacher candidates were observed to deem themselves as moderately self-efficient in terms of the points they scored on the TSEC as a whole (M = 142.56, SS = 40.820) and on its subscales (M_IS = 47.62, SS = 14.169; M_CM = 47.97, SS = 13.885; M_SE = 46.96, SS = 13.873). In addition, PE teacher candidates scored high points on the ASPPET as a whole (M = 83.49, SS = 16.584) and its subscales (M_AP = 48.38, SS = 10.613; M_CP = 35.10, SS = 8.731). The study found no gender-based statistically significant difference between the self-efficacy of PE candidate teachers (t(599) = 1.363, p > .05) and their attitudes toward the PE teaching profession (t(599) = 1.943, p > .05). Another result gleaned from this study was the positive and statistically significant relationship between the points scored on the TSES as a whole and its subscales and on the ASPPET as a whole and its subscales (r = .247, p < .01). The study also showed that PE teacher candidates’ self-efficacy levels did not constitute an important predictor for either the ASPPET as a whole (R² = .062, F = 13.072, p < .01) or for its subscales (AP = R² = .081, F = 17.497; p < .01; CP = R² = .017, F = 3.410 p < .01).