The driving patterns in developing countries, especially in metropolitan cities, are significantly different from those in developed countries or less-populated cities. However, most of the vehicles driven in these cities are tested using driving cycles that do not match those special driving characteristics, typically resulting in an underestimation or overestimation of the emission rates. In this study, a first driving cycle of passenger cars with gasoline engines is proposed for Greater Cairo, Egypt, based on a very large and diverse dataset of high-resolution onboard measurements, recorded using 87 cars driven in 10 distinctive areas in Greater Cairo. Two global driving cycles have been developed based on the commonly used K-means and the newly suggested K-medoids clustering techniques. The developed cycle based on the K-medoids algorithm, which was found more representative of the collected data, is a time series of speed, ~1500 s, covering a distance of ~5.4 km, with an average speed of 12.480 km/h, where 16.3% of the total duration spent in the idling mode. The superiority of the developed cycle over two commonly used cycles, namely the American transient FTP-75 cycle and the European modal ECE cycle, in terms of estimated fuel consumption and emission rates, is reported.