Improve ground cover, root density and organic matter.  Careful selection of native plants and avoiding excessive disturbance to soils can help.  Some plants as legumes may help fix nitrogen to improve productivity.  Adding straw mulch is sometimes helpful in establishing cover.  Avoid using non-native invasive species, but sometimes annual grasses can help establish quick cover, which can help native plants establish.  Careful selection of plant and tree species is needed to help fit circumstances.  Consultation with botanist and soil scientist with restoration experience would be helpful.  If conditions include severe erosion from concentrated flow as rilling or gullying, added practices are often needed to address these conditions.  

Hydromulching can be expensive, but useful if rapid treating roadside or accessible areas with mulch spray, seed, fertilizer as needed.  Chemical additives such as PAM or others may help hold seed and moisture for establishment under some circumstances.  

Refer the following links. The article may be useful. Increasing the inter frictional properties will help in holding sand particles together. This can also be achieved by the inclusion of materials with frictional properties which can prevent sand movement. Also erosion protection films may be used.

http://nanosperse.com/products/sanderosionprotection.html

Using plantation or organic based material may lead to alteration of DO levels in deep streams. I would prefer introduction of natural pebbles for river bed protection against erosion.

In addition to my earlier reply, sand bed streams often lack bed controls, and are susceptable to aggressive changes or avulsion if there are base level shifts or marked changes in sediment loading.  Too much sediment, aggradation.  Too little such as downstream from reservoir or channel sand mining of bed, results in Clean water effect that can cause excessive bank erosion and degradation.  Various structures such as J hooks or cross vanes can sometimes be added to address local bed control and bank stress issues (Rosgen).  Various other structural control, bioengineering and geotechnical approaches have been developed with varying degrees of success.  What is best for a given circumstance is helped by trained and experienced professionals to help evaluate if the issues are localized or due to outside watershed influences that need to be considered.  In many instances, the river will have its way, so try to understand and work with the complexity and forces.  You may be interested in the US NRCS Engineering Handbook Part 650 on stream restoration, available on internet.  It has quite a few examples, photos, etc. that may serve as a helpful reference.  Rosgen wrote one of the chapters.  I co-authored an appendix on gully control.

The question here as I understand is towards controlling the erosion of sand along the stream course due to velocity and down cutting nature of the stream. According to me in such conditions, it is the stream velocity which is to be controlled, as it is difficult to control the erosion of sand under the water by any surficial treatments.

River bank erosion is one of the major and unpredictable hazards worldwide including in Malaysia. Soil detachment at river banks is due to two processes: 1) hydraulic erosion imposed by channel flow and 2) sub aerial erosion due to the weakening and weathering of bank materials. This paper is focused on the second aspect of the erosion process which mainly depends on the combination of rainfall intensity and the ability of the soil to withstand the raindrop effects. The relative combination of sand, silt and clay in a soil is argued to have an impact on erosion resistance. In cohesive soil composition, sand forms the largest size ranging from 0.05 to 2 mm whereas silt is adequately moderate (ranging from 0.002 to 0.05 mm) and clay is the smallest of all three (less than 0.002 mm). With the knowledge that soil composition does indeed have an effect on erosion resistance, this paper will attempt to relate risk assessment index of river bank erosion specifically to soil composition. Thus, the objectives of this document are as follows; 1) to produce risk assessment index for river bank erosion and 2) to carry out a case study for selected rivers in Malaysia pertaining to river bank assessment. The index is produced by inferring the previously developed scale on soil erodibility. Past researchers created the “ROM” scale (named after the researchers, Rolan and Mazidah) to assess degree of soil erodibility into five classes namely “critical”, “very high”, “high”, “medium” and “low”. Instead of using semi empirical formula from the “ROM” scale, a percentage of soil composition was inferred to produce risk assessment index. It was found that as the percentage of clay decreased, susceptibility index became higher and approached a critical level. Application of the newly developed index is verified by conducting a case study at the Langat River, Kajang, Malaysia. The soil composition was classified and form fitted into the index. It was found that the middle reach of the Langat river is susceptible to severe erosion due to low percentage of clay. This finding agreed well with the visual observation of these reaches as a large portion of gully type of erosion had been observed throughout the study. The establishment of risk assessment index which firmly indicates the relationship between soil composition and river bank erosion can be used as a tool in forecasting the risk levels. This formulation is well proven to assess river bank conditions and the associated critical shear stress is very much close with the previously published shear stress.