I might suggest this work by Coleman et al. "Measurement of Multicomponent Solubility Parameters for Graphene Facilitates Solvent Discovery" (Langmuir, 2010, 26 (5), pp 3208–3213 - DOI: 10.1021/la903188a). 

The Authors tested 40 different solvents (see ESI in attachment) and they developed the following protocol to evaluate "graphene" concentration:

" [...]an initial concentration of 0.1 mg/mL of graphite in 10 mL of each solvent was subjected to sonication in an ultrasonic bath (Branson 1510E-MT) for 30 min. To remove large aggregates, the dispersions were then mildly centrifuged at 500 rpm for 90 min (HettichMikro 22R) and the supernatant was retained for analysis. We define the graphene dispersibility, CG, as the concentration of material remaining after centrifugation (CF). We obtain this by measuring the optical absorbance, A, over the wavelength range 200-1000 nm (Cary 6000i UV-vis NIR, l = 1 cm)"

Based on the collected data, they found that "successful solvents are characterized by surface tensions close to 40 mJ/m2", which includes ketones such as cyclopentanone, cyclohexanone and amides such as vinylpyrrolidone, N-methylpyrrolidone, 1,3-Dimethyl-2-imidazolidinone.

The Authors did not test dichlorobenzene, but in another of their works (ACS Nano, 2009, 3 (8), pp 2340–2350 - DOI: 10.1021/nn900493u) they found that using chlorinated solvent such as chloroform and chlorobenzene it is possible to obtain graphene concentration in the range 0.5 - 0.7 mg/mL compared to the best performing solvent they discovered, 1-cyclohexyl-2-pyrrolidone (CHP) which grants concentration of graphene as high as 3.5 mg/mL with the same protocol (see the other ESI file in attachment).

I would personally recommend CHP, but keep in mind that its boiling point is above 330 °C and it is not easily removed. Nevertherless, I used it with excellent results to characterize and functionalize CNTs in solution (https://www.researchgate.net/publication/51466112_The_continuous-flow_cycloaddition_of_azomethine_ylides_to_carbon_nanotubes?ev=prf_pub)

Article The continuous-flow cycloaddition of azomethine ylides to ca...

Have a look

http://www.sciencedirect.com/science/article/pii/S0021979714003397

Thank you brother for the advices and link of papers. I just wondering one thing, if a polymer will be mixed with graphene in order to construct polymer-graphene composite, why people do not mix it directly, that is the graphene with polymer. in fact, they functionalize the Graphene Oxide first, later on mix it with the polymer. Thank you

First, let me suggest the paper in attachment for a proper classification of graphene-like materials. As stated therein: "One concern in graphene research is that the term graphene is used in a generic manner to describe many graphene-based materials (GBMs)."

Now, depending on each specific GBM, there might be different approaches to the preparation of carbon-based nanocomposites. Arguably,

a) the economically advantageous solution could be what you are suggesting: process micrometric flakes of graphite, partially exfoliated graphite (aka graphite microplates, few layer graphite, few layer graphene) or reduced graphene oxide (aka rGO, graphite nanoplates, graphene nanoplatelets) in an extrusion process to produce a nanocomposite masterbach, like the Nanocyl carbon nanotubes masterbatches. The advantage is that no solvents are required in an extrusion processes since the dispersion of the nanostructures in the polymer is induced by the mechanical work of the extrusion screw. Moreover, you can buy cheap rGO from many commercial sources (I have personally work on the rGO materials supplied by Graphenea and ACS Materials).

b) another option could be to prepare polymer nanocomposites by solution-based processes. Unfortunately, most of the GBMs discussed above have limited solubilities in common solvents. Therefore, the use of soluble graphene oxide or functionalized graphene-derivatives. The difference between the two is that graphene oxide derivatives have to be reduced after compatibiliazion in the polymer to obtain graphene-like materials, while functionalized graphene-derivatives do not require post-treatment usually.

c) the inconvenient and expensive approach would be to growth a perfect monolayer graphene by CVD and try to use it a filler in a nanocomposites.

Let me conclude suggesting a recent review on "graphene" nanocomposites by Vika Mittal (http://onlinelibrary.wiley.com/doi/10.1002/mame.201300394/abstract) and, if you are interested in functionalization processes, you might be interested in my work on the decoration of carbon nanostructures (including nanotubes and graphite derivatives) by a RAFT polymerization initator (http://pubs.rsc.org/en/content/articlehtml/2013/cc/c3cc44700f).

I agree with some good points from Patrizio. First, you had better clarify which type of your graphene sample is: pristine graphene or rGO; Second, precisely speaking no solvent can dissolove graphene (even in the form rGO), but disperse it at best.