Abstract Particle emissions from motor vehicles have been associated with local air quality issues and human health concerns. Epidemiological studies have indicated exposure to ultrafine particles, measured as particle number as opposed to mass, are linked to cardiovascular and respiratory diseases. Although hybrid-electric vehicles (HEV) have an increasing presence in the on road fleet and have been designed to minimize emissions and energy consumption, little has been done to quantify the real-world tailpipe emissions from these vehicles. Second-by-second data were collected from the tailpipe of a Toyota 2010 Camry hybrid and a conventional vehicle (CV) model equivalent along a 32-mile route in Vermont over an 18-month study period. The vehicles were instrumented with the Total Onboard Tailpipe Emissions Measurement System (TOTEMS), capable of measuring particle number size distributions (TSI EEPS), exhaust flow rate (custom tailpipe adapter), spatial information (Garmin and Geologger GPS), and vehicle engine control unit information (Toyota Techstream Scantool). The HEV, on average, had higher particle number emission rates than the CV, particularly for the urban portion of the route where the HEV was expected to demonstrate improvements over the CV. It has been observed that the HEV internal combustion engine restarts produce peak emissions four times that of stable HEV engine-on operation and 3.5 times that of stable operation in the CV. With many of these restart events occurring in areas with frequent stop-and-go and in close proximity to human activity and pedestrians, closer examination was necessary. The goal of this work was to predict the particle size distributions at HEV engine restart events. Particle size distributions at engine restart were predicted as a function of engine-off time preceding a restart, motor torque, generator torque, vehicle speed, and calculated load. The ability to predict these high emission events may allow for adjustments to the control algorithms of the HEV drivetrain to further prevent restart particle emissions and potential human exposures.