If you identify a center of origin, a priori, you are assuming dispersal from a center of origin, that is, you dismiss vicariance, and there is no reason to do so.

The ancestral area of a species complex with four species, A, B, C and D, may be inferred to be the area occupied by all four together.  The relationships among these areas, as inferred from a phylogeny of the four species, is analogous to one character-state tree in a phylogenetic analysis.  You need to compare it to the relationships of the areas as specified by other taxa in the biota in order to hypothesize a set of relationships for all of the areas. 

What kind of species complex you mean? Morphological? Evolutionary? Ecological?

You might want to consult this publication by Nylinder et al. They used a diffusion model to reconstruct ancestral areas for a set of Centipeda species.

https://www.researchgate.net/publication/259320373_On_the_Biogeography_of_Centipeda_A_Species-Tree_Diffusion_Approach

Article On the Biogeography of Centipeda: A Species-Tree Diffusion Approach

It does depends on the data you have. If, for example, you have a single-locus phylogenetic tree, you might want to try the Lemmon and Lemmon (2007) likelihood method.

Article A Likelihood Framework for Estimating Phylogeographic Histor...

As it has been mentioned in several other questions; it depends first on your organism; then take a look to the biogeography of the whole group and higher taxa and it will start to give you some lights

I support the opinion of Eduardo: You have to study biogeography of the whole group and its higher taxa, plus - paleontology. Sometimes, the species origin is linked to a single geological/climate change event in remote era.     

If you identify a center of origin, a priori, you are assuming dispersal from a center of origin, that is, you dismiss vicariance, and there is no reason to do so.

The ancestral area of a species complex with four species, A, B, C and D, may be inferred to be the area occupied by all four together.  The relationships among these areas, as inferred from a phylogeny of the four species, is analogous to one character-state tree in a phylogenetic analysis.  You need to compare it to the relationships of the areas as specified by other taxa in the biota in order to hypothesize a set of relationships for all of the areas. 

The fossil record can help.

As everybody says it depends on your data, but you would like to check the range expansion hypothesis in MS as Slatkin 1993, Slatkin and Excoffier 2012, Débarre et al 2013.

This may help. 

Yu, Y., Harris, A. J. & He, X. J. (2010). S-DIVA (Statistical Dispersal-Vicariance Analysis): A tool for inferring biogeographic histories. Molecular Phylogenetics and Evolution, 56(2), 848-850.

Yu, Y., Harris, A. J. & He, X. J. (2012). RASP (Reconstruct Ancestral State in Phylogenies) [v.1.107]. Available from http://mnh.scu.edu.cn/soft/blog/RASP.

If you map the geographies onto a tree,I guess the regions of the species in most basal clade would give you an idea about the place of origin. 

What is "the most basal clade"? All clades have a sister clade that is equally basal. Even if you define a basal clade as "the least diverse",  the method would not work, as relict lineages need not necessarily remain the certer of origin. The method of mapping geography on a tree might work only of if geography is optimized unambigously on the root node of the tree, but the root is generally not refered to as a "basal clade".

 I agree Prof. Paulo. That's a lot of information which helps my studies as well. Of course just by looking at the tree topology we can't come to a conclusion. But according to what I have learned, we can get a  basic idea about the most ancestral taxa by looking at the phylogenetic tree.  Hope this would help Xinxin.

http://www.sciencedirect.com/science/article/pii/S105579030300438X