Big caveat up front: many people struggle A LOT in their early graduate school career, and this doesn't mean that you can't be successful. A better guide to whether you should switch to theory might be whether or not you find the day-to-day process of debugging your experiments satisfying - even when it fails a lot of the time. (Also, if you like programming much more than being at the bench, that's another good sign for switching.)
If you are early in your career (first couple of years of graduate school), switching may be relatively easy, especially if you want to stay in the same university and department/program. I certainly knew many students who changed advisors early, though relatively few switching theory/experiment.
In general, I would suggest looking at prospective advisors, and talking with their current students to see what you'd need to know, and finding out if they would take students. Classes are useful, but most advisors will have their own recommendations for which are most important. Even beyond your own skills, other factors like research funding, will be critical in which advisors can accept you. Also, personal fit with an advisor is just as important as research area. Because of these points, it is probably better to figure out what group you want to join first, and then see what courses you should take.
I would also recommend that, if you have strong computational skills (programming) and a good physics/analytical methods background (i.e. good grades in core courses or other evidence of this), and are interested in biophysics, there is no need to limit yourself to systems biology / biological networks. There are a lot of computational biophysics topics out there, ranging from bioinformatics, protein interaction networks, computational neuroscience, biomechanics, molecular dynamics, etc., and if you have evidence of flexibility and learning a lot of different skills, you could be considered by groups focusing on topics further away from your past experience than you think.
All that out of the way: for a biophysicist with an interest in theory, other courses that are worth it are things like statistical mechanics, especially something focusing on non-equilibrium statistical mechanics. That could be in chemistry/chemE departments. Within network fields (which I know less well), you might want to infer network structure from data, which means there may be some statistics or machine learning courses that are worthwhile. For biomechanics-related theory, a continuum mechanics class (elasticity theory/hydrodynamics/etc) might be worthwhile.