The great majority of overhead distribution systems in urban and rural areas of countries such as Australia, Iran and etc., are in interfering with vegetation. This phenomena leads to high impedance faults (HIFs) which not produce enough fault current to be detectable and consequently causes electrical energy losses. In this paper, according to experimental data obtained from measurements, the effective factors in power losses caused by trees are studied and a new numerical model of power loss is presented so as to reflect the effects of environmental conditions and biological classification. The base of this method is according to Samuelson theory. In proposed algorithm, the impacts of species, short-term environmental conditions (temperature and humidity) and long-term environmental conditions (Seasonal variations in physiology) on the power loss estimation are fully considered. Experimental investigations on a real low voltage (380 V) and medium voltage (20 kV) distribution network verify the algorithm’s operation.