Generally, nanoparticles are stable in the suspension if they have a sufficiently strong barrier that prevents aggregation. In chitosan's case perhaps the positive charge on the surface will make particles repel each other. You should also realize that the characteristics of the buffer (pH and ionic strength) would influence the interaction forces between the nanoparticles. Therefore, it may be a good idea to make sure that your nanoparticles have the same characteristics (including charge) and that you use exactly similar buffers (pH and ionic strength). Also make sure that your buffer and particle characteristics are not influenced by temperature. It is noteworthy that if the suspension is stabilized by charges on the nanoparticles then it is normally more stable in low ionic strength buffers and at pHs far from the isoelectric point of the charged groups of the nanoparticles. So that is something you can try. If all that is taken care of  and you still get aggregation at 25 (and not at 5) it probably is a kinetic issue meaning that your particles like to aggregate (the repulsive barrier is not strong enough to keep them singlets) and the only reason that they do not aggregate at low temp is because they do not collide that frequently whereas at 25 their movement results in high collision efficiency and aggregation.  I would try to increase the positive charge on the nanoparticles or play with the buffer characteristics (low ionic strength and optimal pH) to get a stable suspension.

I hope this helps.

 Generally, nanoparticles are stable in the suspension if they have a sufficiently strong barrier that prevents aggregation. In chitosan's case perhaps the positive charge on the surface will make particles repel each other. You should also realize that the characteristics of the buffer (pH and ionic strength) would influence the interaction forces between the nanoparticles. Therefore, it may be a good idea to make sure that your nanoparticles have the same characteristics (including charge) and that you use exactly similar buffers (pH and ionic strength). Also make sure that your buffer and particle characteristics are not influenced by temperature. It is noteworthy that if the suspension is stabilized by charges on the nanoparticles then it is normally more stable in low ionic strength buffers and at pHs far from the isoelectric point of the charged groups of the nanoparticles. So that is something you can try. If all that is taken care of  and you still get aggregation at 25 (and not at 5) it probably is a kinetic issue meaning that your particles like to aggregate (the repulsive barrier is not strong enough to keep them singlets) and the only reason that they do not aggregate at low temp is because they do not collide that frequently whereas at 25 their movement results in high collision efficiency and aggregation.  I would try to increase the positive charge on the nanoparticles or play with the buffer characteristics (low ionic strength and optimal pH) to get a stable suspension.

I hope this helps.

I would agree that the best thing is to maybe adapt your buffer characteristics or if this is possible for your research approach increase the positive charge on the nanoparticles.

What I do suggest to you is to measure the zeta-potential of your colloidal solution at both temperatures. This helps you to understand the mechanism of stability or instability of chitosan nanoparticles. If the magnitude of zeta potential at 2C and 25C is larger and smaller than 30 mV, respectively, then the mechanism of stability is electrostatic repulsion. In this case, tuning pH and not ionic strength might be useful. If in both temperatures the magnitude of zeta potential is smaller than 30 mV, then the mechanism of stability at 2C might be by steric repulsion. In this case, you must do an independent study on the chitosan surface charge at different pH values and ionic strengths. This helps you to find the conditions at which chitosan molecules would carry electric charge helping the nanoparticle stabilization. 

Adding a bit to the previous explanations and suggestions, you could modify the properties of the particles either by adding charge (using PEI for example), or by eliminating H-bond donor groups, for example by Pegylating the surface (using methoxy terminated TosOPEG).

You have to control pH. An acid environment generate aggregation.

Concentration is another issue. A higher concentration favor self-assembling.

Nature of solvent is another important aspect. CHCl3, for example is favoring aggregation.

Navid Bizmark gave to you valuable information and you can use some surface agents to prevent aggregation.

I would have doubts that zeta potential at the 2 temperatures is much different.

It maybe that steric stabilization is reduced at higher temperatures. Besides increasing electrostatic and steric stabilization, i.e. increasing repulsion you can weaken the van der Waals attraction by matching refractive index of dispersion media and dispersed particles.