the answer depends on several factors: material, state (aggregate?), crystalline/amorphous and on the type of size you want to obtain (look for several discussions on this on the site)..
If the system is crystalline then you can definitely try powder diffraction. It is fast and the result corresponds to a statistically relevant quantity of specimen. Plus you don't need a dedicated SAXS machine, you have a sensitivity to lower moments of the size distribution (3rd), you don't need a diluted specimen and you have the directional information from which you can understand the shape features. You obtain however the so called "domain size" that corresponds to the grain or particle size when you deal with well behaving powders
I agree with Andreas that DLS is an easy technique for suspended nanoparticles. Nanoparticle Tracking Analysis (NTA) is also gaining ground with additional benefit of being able to "visualize" the particles and getting average size not biased for larger particles. See this page: http://www.nanosight.com/technology/nanoparticle-tracking-analysis-nta
I agree also with Andreas, DLS is the faster technic to estimate size of nanoparticles but results are more complicated to interpret if particles are not spherical
Well i agree with adreas too, But DLS (Dynamic Light Scattering) is not a precise technique to give real value of size of nanoparticle since you want too study the dispersion more than the size of nanoparticle which in dispersion medium is the hydrodynamic radius depends on many factors of dispersion medium. With DLS as a technique to understand the stability of the Colloidal dispersion medium you should also study Zeta potential. Also, Polydispersity Index can be calculated from DLS technique for concluding the polydisperse nature of the colloidal dispersion but concluding that the Nanoparticles are Spherical is not concluded on strong basis using this technique. This both technique would give you a better picture of the Colloidal dispersion stability and also the hydrodynamic radius of the Nanoparticle in the dispersion medium.
To know precise size and shape of nanoparticles TEM is the best technique, To know more details of Nanoparticle XRD analysis would be better, Few nanoparticle would also follow Raman Spectroscopy Studies and magnetic details could be found out from magnetic hysteresis technique and VSM.
Agree that DLS might be used to get a first info about differences in average size. Zeta potential might give a hint for electrostatically stabilized systems, only for those. NS information might be misleading nevertheless in the nanosize range. See e.g.
Evaluation of particle interactions by in situ visualization of separation behaviour
Particle size distribution by a method based on particle separation can give you much more precise info about fraction of agglomerates and distribution shape. Even without doing analysis of size distributions separation behaviour gives you much information about the dispersion state of your particles and the state of the formulation.
I agree that the DLS is a good method to determine the size distribution of nanoparticles, provided that you have previously defined their form by other methods. However, most conventional DLS instruments uses an algorithm based on the Mie theory, ie, for spherical particles (see http://www.malvern.com/LabEng/products/zetasizer/zetasizer.htm). If you have a different particle shape, for example, a cylinder (tobacco mosaic virus) you can get any particle size. The problem can be solved by using angular measurements of light scattering.