Kindly check here..

http://www.crs4.it/repository/documents/conferences/swat_2015/introductory_manual_updated_oct._2012.pdf

Check this too..

http://biomath.ugent.be/~ann/ArcSWAT_Nzoia.pdf

Also refer this web info..

http://www2.fiu.edu/~melessea/publications/swat1.pdf

See here too..

https://web.ics.purdue.edu/~vmerwade/education/arcswat.pdf

Refer this too..

https://groups.google.com/forum/#!topic/swatuser/eVOEg0DEM-s

Small correction in this https://web.ics.purdue.edu/~vmerwade/education/arcswat.pdf while considering slope in HRU analysis. as HRU's represent unique response unit, which consists only one soil, LULC class, it is recommended to consider multi classes of slope instead of one. It may helps to route the water from each sub-watershed to basin outlet.

Thanks for the valuable answers..:)

Refer the below paper.

Article Streamflow and Sediment Yield Prediction for Watershed Prior...