A five-stage prediction-observation-explanation inquiry-based learning (FPOEIL) model was developed to improve students' scientific learning performance. In order to intensify the science learning effect, the repertory grid technology-assisted learning (RGTL) approach and the collaborative learning (CL) approach were utilized. A quasi-experimental design study was conducted to examine whether the students who used the FPOEIL model only had better learning performances than those who used FPOEIL with RGTL or CL. This study adopted purposive sampling, selecting 123 fourth grade students. The experimental process was conducted during five weeks. It was found that the FPOEIL model improved the students' learning performance. Moreover, the low prior knowledge students who learned science using FPOEIL with RGTL or CL had better learning performances than those who learned using the FPOEIL model only, and the effectiveness showed no significant differences between the low prior knowledge students and the high prior knowledge students. Using the FPOEIL model, the positive effects were intensified in the continuous inquiry-based learning activities and feedback-correction process for the students learning science. The RGTL approach helped the students find, remember, and comprehend scientific knowledge. In the CL process, the students spent more time discussing how to integrate clues to answer the science question.