Actually, according to the Rayleight formula the bes angle should be close to zero. The sygnal should be about twofold higher than at 90 degrees. As far as I know in dynamic light scattering the angle 8 degrees is used. 

The theoretical description could be done for relatively simple colloidal partciles. Their non-spherical shape or aggregation typically make the picture extremely complex. Than it is essential to take into account refraction at the cuvette walls properly.  

Iftak,   It depends on the size of suspended particle.   If the particle is small enough such as order of wavelength or less and can be well approximated by sphere, mie scattering theory may be successfuly applied.   In term of angle dependent scattering, mie scattering thery tells 90 degree is not best direction for detection but not too bad.    Shigeo

@ Shigeo Kubota .....sir, if we consider clay particles, which is the cause of water turbidity, the shape and size of these particles are largely varying from each other or the particle distribution is too broad. Therefore in determining turbidity it is prescribed by EPA to use a standard suspension, Formazine standard and compared the intensity of the turbid water to this standard and expressed the turbidity as FNU (Formazine nephelometric Unit). I am using this method, it gives good results as compared to the standard turbidimeter but as a physics student I need to see the theoretical aspect of this study. But, unfortunately I am not able to find any theory till now.  

Iftak,   If average clay particle is much larger than visible wavelength, scattering crosssection enhancement by diffraction effect can be neglected.   In such geometrical optics limit, simple Lambert Beer's rule is applied.   If you want to treat it by Mie or Rayleigh scatterings, at least particle size information as well as its complex refractive index is required.   Mie scattering is applied for a wide variety of applications such as atmosherics Tyndall phoenomena, colloidal particles, blood corpuscles..  Shigeo