Abstract The valet parking system is made up of carefully built yet simple-to-use softwareapps that allow even the most technologically challenged persons to access ride-on-demandservices. As a result of our findings, we now realise the necessity for an on-demand valet servicesystem in India's metropolitan centres, where the population is growing and car traffic isexpanding. This project creates an architecture based on the Firebase Cloud Messaging (FCM)technology and introduces a novel algorithm that provides a valet parking system. This costwill be used to provide a solution for identifying an available valet service provider in responseto a user's request, as well as a solution for proposing a new valet service provider if the presentvalet service provider refuses the request or no one is available to serve at the time. Thesimulation findings suggest that the algorithm assists users in receiving on-demand valetservice and reduces the time spent looking for parking in real time. We will also be successfulin putting the planned system into practice.Keywords: car parking, valet, machine learning, k nearest neighbor, android, java, etc