The purpose of this study was to investigate the relationship between students’ attitudes toward programming, gender, and learning performances. The survey used for measuring students’ attitudes toward programming consisted of 20 questions on a five-point Likert scale in five dimensions (meaningfulness, interest in programming, self-efficacy, creativity, and collaboration). Ninety freshmen who had basic programming experience by using block-based programming in the Innovation in Educational Technology course were asked to take the survey. The overall reliability of the survey was found to be 0.93. The results showed that there was no significant difference between male and female freshmen in attitude toward programming, but there was a significant difference among different learning performances in dimensions of interest in programming, self-efficacy, and creativity. We performed pairwise comparisons at the same level of significance by using Fisher’s least significant difference (LSD) method to test which group differs from the other groups. The results found that low-performing students’ attitudes toward programming in dimensions of interest in programming, self-efficacy, and creativity were the lowest of all types of students. This is a challenge for instructors in planning learning activities to encourage low-performing students to have a more positive attitude toward programming.