The invention relates to an optimal distribution method for a thermal power generating unit load based on a dynamic inertia weighted particle swarm. The method comprises the steps of constructing a function relation between all unit loads and a standard coal consumption amount by a thermodynamics test; fitting a coal consumption characteristic curve according to coal consumption characteristic parameters of a generator unit, constructing an optimal load distribution model, converting the optimal distribution into a non-constraint problem by a penalty function method on the basis of influence of various types of constraint conditions, different fuels and different operation limitation conditions during operation of a power plant generator, and distributing main generating power to all units to finally obtain an optimal load distribution result for all the units by combining a dynamic inertia weighted particle swarm algorithm. Therefore the coal consumption of a thermal power generating unit is reduced; on the premise that hardware facilities of a power plant is not modified, a load instruction of each unit is only needed to be redistributed, so that the economy of the operation of the power plant can be effectively improved, and comprehensive energy utilization, energy conservation and emission reduction are guaranteed; the method has an active significance for propelling energy conservation and emission reduction of a country.