Venoms of wandering or running spiders from the family Ctenidae are multicomponent and include 500 identified peptides and proteins called ctenitoxins. The main components of the venom are cysteine-rich peptides containing an inhibitor cystine knot (ICK) motif. The pharmacological diversity of ctenitoxins makes it possible to consider some of them as prototypes for the development of new drugs to treat chronic pain, Huntington’s disease, erectile dysfunction, and glaucoma. Based on the localization of cysteine residues in the amino acid sequence, ctenitoxins are divided into 14 groups containing 6–14 Cys residues. To date, the spatial structure of only one ctenitoxin, ω-CNTX-Pn4a (Pha1β or Tx3-6) from the Brazilian wandering spider Phoneutria nigriventer, has been determined. Another ten structural groups of ctenitoxins are homologous to the known spatial structures of toxins from other spider families or other proteins; finally, three more ctenitoxin groups do not have any known structural homologues. In this study, we proposed possible patterns for the disulfide bond formation in all ctenitoxin groups. A comparison of the obtained schemes with the predictions of the AlphaFold 2.0 algorithm shows that this neural network does not always correctly predict the structure of cysteine-rich peptides, especially in the case of modeling of mature peptides without leader sequences.