This paper focuses on the design of project images on a crowdfunding website, which portray the themes and contents of the projects. Employing the Stimulus-Organism-Response (S-O-R) model, we investigate the relationships between image attributes (S) and image emotions (O), and between image emotions (O) and campaign outcomes (R). We develop and train a deep neural network machine learning model to identify the emotions conveyed in the images, and then implement it to project images from a popular crowdfunding platform. We apply the objective measurements of image attributes, the obtained image emotions, and the project outcome metrics to the S-O-R model to explore how image emotions are related to the success of crowdfunding projects, and from a design perspective, what image attributes evoke the image emotions. In an extended study, we conduct an online randomized controlled experiment by manipulating image attributes. The experiment results also show that image attributes can influence emotions and support that participants’ emotions are related to their pledge intention. This research contributes to the crowdfunding literature and practice, and machine learning in image analysis.