The rise in popularity of ride-sharing services and ride-booking systems has created new opportunities and challenges for security and safety. A useful system for passenger safety assistance using machine learning and mobile applications is missing from the existing work. This paper develops a data set regarding suspicious activity detection using a questionnaire. This paper selects a suitable machine learning model for suspicious activity prediction during a transport ride by examining support vector classifiers (SVC), random forest, MLP classifiers, decision trees, KNN, logistic regression, and Gaussian naive Bayes classifiers. The results showed that the SVC is most suitable, with 97% accuracy, for classifying suspicious activity predictions during transport riding. This paper provides a passenger safety mobile application with passenger and driver verification, application rating, suspicious activity prediction, suggestions regarding safety, location mapping, and trip booking features. The application evaluation results based on users’ comments showed that more than 55 percent of users supported the application's usability and effectiveness nature.