Introduction: Exergames is defined as a technology-driven physical activity, which is an innovative way of physical activity that integrates interactive gameplay in the exercise process. The exergames may provide enjoyable experiences that could motivate people to participate and continue playing the game play, while also exercising at the same time. Methods: This article presents a treasure hunt-based walking exergames on android platform with the implementation of intelligence-based image recognition. The exergame, termed USM ExerHunt uses images of Universiti Sains Malaysia buildings as the hints. The participant of the game supposes to find a building shown in the hint, and once reaching the destination captures the image of the building. Then, the application will calculate the total steps taken and calories burnt by the participant using an implementation of accelerometer from the mobile phone. Results: The developed USM ExerHunt application is able to achieve accurate image recognition of USM building, with the accuracy rate of 92%. Besides that, the application is capable of calculating the number of total steps and calories burnt after an exercise routine is completed. Conclusion: This android application has shown a proof of concept in incorporating machine intelligence into an exergame application, with pilot study within the USM community.