In order to simulate the real behavior of vehicles during laboratory tests, such as fuel consumption  and pollutant emissions , coast down coefficients must be set on the chassis dynamometer control system. These coefficients are used to determine load curves, which represent the resistance imposed on the vehicle movement by the wheels rolling and the air, being obtained from track tests performed according to Brazilian standard ABNT NBR 10312 . However, coast down tests depend on the availability of long and flat tracks. This may entail costs for deployment or leasing of facilities with these characteristics, which may include even long commutes of human and material resources, depending on its location. This paper proposes an alternative methodology for coast down coefficients determination, from experiments on chassis dynamometer and vehicles aerodynamic specifications. It was applied to some Petrobras Research Center (Cenpes) test vehicles. The resulting coefficients and load curves were compared to those obtained on the track. Furthermore, wheels rolling and air resistances are shown separately, which is not trivially achieved in track tests. This allowed an analysis of the load curve adjustment criteria adopted in the Brazilian norm.