This paper presents a general intelligent optimization numerical model to identify multi-variables of the non-linear one dimensional inverse heat conduction problem in transient state by a precise algorithm for direct heat conduction, based on Ant Colony Algorithm. The inverse problem is formulated implicitly as an optimization problem with the cost functional of squared residues between calculated and measured quantities. Combined identifications can be carried out for non-linear source term, thermal diffusivity and boundary conditions etc. Satisfactory numerical validation is given including a preliminary investigation of effect of noise data on the results. Results show that the proposed numerical model can identify combined parameters for the inverse heat conduction problems with precision.