A reliable outcome measure is crucial to understand the impact of emerging interventions, such as animal assisted therapy (AAT) for individuals diagnosed with Autism Spectrum Disorder (ASD). A video analysis tool was created to measure the impact of AAT on people with ASD by recording the frequency and duration of positive, play and negative behaviours. This study aimed to determine the inter-rater and test re-test reliability of the video analysis tool and to investigate if session complexity and eater factors influenced the reliability of the tool in this population. Following training in the use of the tool, 23 occupational therapy students from an Australian metropolitan university rated a simple or complex video-recorded AAT session on two occasions. The raters also completed a survey to obtain demographic information. Expert raters determined acceptable score ranges which were compared with collected data in SPSS to determine Intraclaas Correlation Coefficients ( ICC). The ICCs of the video analysis tool were 0.84 (simple session) and 0.894 (complex session ) for inter-rather and 0.84 (simple session) and 0.89 (complex session) for test re-test reliability. The percentage agreement was similar regardless of session complexity and rater experience with children and ASD but varied across subscales and rater animal experience. This study indicates the video analysis tool has potential as a reliable outcome measure for AAT with individuals with ASD, however further research utilising increased sample sizes is required to establish the effectiveness of the tool.