I am currently doing DESeq2 analysis, and I am getting opposite log2 fold change when I plot on Volcano plot.

For example, one of my genes is down-regulated (with significant p-value) according to DESeq2; however, it says otherwise when I look at the actual gene expression data.

What am I doing wrong? Thanks for your help.

Take a look at the vignette note on factor levels, and also see what is printed at the top of your results table when you print it in the console:

> res

This is an indicator for the user to know what R used as the reference level if you did not set it yourself (again see vignette note).