Visual representation is crucial for enhancing students' conceptual understanding of fractions and correcting misconceptions. This explanatory-sequential mixed study assessed Indonesian preservice mathematics teachers' perceptions, self-efficacy, and implementation competencies using visual representations in fraction learning and identified gaps among these three aspects. A total of 251 students from three study programs participated in a validated online survey (α = 0.89). Quantitative data were analyzed using descriptive statistics. Seven participants were then purposefully selected for semi-structured interviews; transcripts were analyzed using Braun & Clarke's six-stage thematic analysis with researcher triangulation. The results showed that although prospective teachers had a very high perception of the pedagogical value of visual representations (M = 4.41; SD = 0.153; 78.9% agreed/strongly agreed), their self-efficacy level was only in the moderate category (M = 3.21; SD = 0.617), with only 28.3% showing high efficacy, and 61.3% were in the medium category. Implementation competence was even lower, with only 17% of participants being able to select and explain visual models pedagogically and 27.8% meeting the sufficient criteria in the teaching reasoning test. The combined analysis of quantitative and qualitative data confirmed the existence of a systemic gap between strong conceptual understanding and limited practical skills, which was mainly caused by uncertainty in model selection, minimal misconception-oriented practical experience, and weak pedagogical reflection. Integrating high-fidelity micro-teaching, representational scenario-based simulations, and video reflections is recommended to enhance self-efficacy and implementation competency.Visual representation is crucial for enhancing students' conceptual understanding of fractions and correcting misconceptions.This explanatory-sequential mixed study assessed Indonesian preservice mathematics teachers' perceptions, self-efficacy, and implementation competencies using visual representations in fraction learning and identified gaps among these three aspects. A total of 251 students from three study programs participated in a validated online survey (α = 0.89). Quantitative data were analyzed using descriptive statistics. Seven participants were then purposefully selected for semi-structured interviews; transcripts were analyzed using Braun & Clarke's six-stage thematic analysis with researcher triangulation.The results showed that although prospective teachers had a very high perception of the pedagogical value of visual representations (M = 4.41; SD = 0.153; 78.9% agreed/strongly agreed), their self-efficacy level was only in the moderate category (M = 3.21; SD = 0.617), with only 28.3% showing high efficacy, and 61.3% were in the medium category. Implementation competence was even lower, with only 17% of participants being able to select and explain visual models pedagogically and 27.8% meeting the sufficient criteria in the teaching reasoning test. The combined analysis of quantitative and qualitative data confirmed the existence of a systemic gap between strong conceptual understanding and limited practical skills, which was mainly caused by uncertainty in model selection, minimal misconception-oriented practical experience, and weak pedagogical reflection. Integrating high-fidelity micro-teaching, representational scenario-based simulations, and video reflections is recommended to enhance self-efficacy and implementation competency.