width treatment

Hello, my question may seem silly and I'm grateful for your patient reading and replying.
I'm studying some exotic higgs decay processes and want to make sure that my treatment of width/branching ratios is correct. Specifically, I start from the SM model file supplied by FeynRules and add a new light scalar particle a1 and trilinear interaction ha1a1 to creat a new model in order to study exotic higgs decay h > a1 a1. The SM model supplied by FeynRules has set higgs width to be 0.00407GeV (which is smaller than the partial width of h > b b~ calculated from bottom Yukawa in the model at tree level). Now I want to generate processes for a benchmark point which has Br(h > a1 a1)=0.1. My way of achieving this is tuning trilinear ha1a1 coupling and/or higgs total width in feynrules model such that the partial width of h > a1 a1 calculated by MadGraph divided by the total width I set in FeynRules is just equal to 0.1. (Assuming the value of total width doesn't spoil narrow width approximation.) Then I use the FeynRules model with the tuned coupling/total width parameter to study processes like p p > w+ h, h > a1 a1, etc.
I wonder whether my treatment is correct, especially when the higgs total width I set is smaller than h > b b~ partial width.
By the way, if I'm not studying h > b b~ decay, will the fact that the higgs total width I set is smaller than h > b b~ partial width cause any potential problems concerning cross section calculation or event generation for other processes?
Thank you very much!