Dear Lan --
I usually don't read the output text files directly, and if I do, I use R to read in the files and manipulate the data.
Instead, I usually use the ArcSWAT interface to load the text files into an Access database, and then within Access I extract the output I want from the tables in the database. To do this, after running SWAT, choose SWAT simulation/Read SWAT Output. Choose the output files to import into the database (I usually choose output.rch, and sometimes output.sub, output.hru, and output.rsv). Then click "Import Files to Database."
The output database is in the Scenarios/Default/TablesOut folder of your project database. It is called "SWATOutput.mdb."
Then I use queries to extract the data I want. It's beyond the scope of this forum to teach users how to use Access queries, but start with simple "select" queries to get the data (for example) for selected reaches and selected years. You can aggregate data to get averages or sums with the "Totals" option. With a few auxiliary tables giving more user information about each HRU, subbasin, or reach, you can extract and summarize data in more sophisiticated ways by joining these tables to the SWAT output tables. I almost always create an HRUinfo table that identifies further information about my HRUs, like what crop rotation is in each one, which HRUs are grazed, what the hydrologic soil group is, and so forth. I also create a SUBinfo table identifying which subbasins are in a region of interest (usually tributary subwatersheds).
In short -- it is worth learning how to write queries in Access to extract the data of interest from SWAT output tables. Understanding a little about database queries is a useful skill for most scientists and engineers.
Then I often copy and paste from Access into Excel for simple plotting during calibration (I do all my final plots in R).
-- Jim