This study aims to describe the effectiveness of the Participatory Observation learning model and students’ responses to the learning model used in PAI (Pendidikan Agama Islam/Islamic Education). It used qualitative and quantitative methods with a pre-test and post-test nonrandomized group pre-experiment design. The 125 participants who were taken from three public universities, were asked to fill out questionnaires and written interviews. The data were analysed using paired t-test and qualitative descriptive analysis. The results showed that first experiment got a t-score of 6,798 with a significance of 0,000; t-table score at 51 degrees of freedom and a significant degree of 0.05 in the form of 2.032. Second experiment obtained a t-score of 5.991 with a significance of 0.000; the t-table score at 54 degrees of freedom and 0.05 degrees of significance was 2.046. Third experiment achieved a t-score of 6.424 with a significance of 0.000; the t-table score at 20 degrees of freedom and 0.05 degrees of significance was 2.054. When compared, t-score> t-table or significance score <0.05, which means that the mean score between pre-test and post-test is significantly different. This means that the Participatory Observation learning model can improve students' religious and social character in learning PAI. This result is in line with the respondents' recognition that they have experienced changes in their character in a positive direction. Students stated that the Participatory Observation learning model was fun and very useful, so that it could be used in PAI learning to improve students’ character.