I need to know which method is the best for modling antennas, that have some active components as pin diode or varactor diode? and their feeding circuits, are they affecting the modeling?
When you have active devices at the feed, particularly with nonlinearities, it is often better to model the antenna to the feed point with a frequency sweep or transient response and then to interact that result with the feed network in a nonlinear circuit analysis program. Once you have to resultant feed current with the network, you can then return to your antenna analysis to get radiation and such. If you are writing your own, I would suggest MoM (either in frequency or time) using the EFIE, though FEM or FDTD should do the job also. If you are using a commercial package, they should all give similar answers. With FEM and FDTD, you have to develop radiation boundaries. MoM will require the termination of the ground plane (so do the others).
If you are writing your own, a lot depends on how well you use the method. That would include basis functions for MoM, discretization methods for integrals and such, the sample size (too small vs too large) and such. Most commercial packages have already addressed these aspects to get good results, but folks writing there own code often ignore some fundamentals and then cuss out the method.
The first answer is compliant and it depends if you want to use commercial software or you want to built yours. the FEM (HFSS) , MoM (FEKO) are more used . I advise you to use them to valid your results because one method use differential equations techniques and the second use the integral equations techniques. if you want to built yours it depend at what part od device you give more interest. have a nice day