With the urgent need for new medical approaches due to increased bacterial resistance to antibiotics, antimicrobial peptides (AMPs) have been considered as potential treatments for infections. Experiments indicate that combinations of several types of AMPs might be even more effective at inhibiting bacterial growth with reduced toxicity and a lower likelihood of inducing bacterial resistance. The molecular mechanisms of AMP-AMP synergistic antimicrobial activity, however, remain not well understood. Here, we present a theoretical approach that allows us to relate the physicochemical properties of AMPs and their antimicrobial cooperativity. It utilizes correlation and bioinformatics analysis. A concept of physicochemical similarity is introduced, and it is found that less similar AMPs with respect to certain physicochemical properties lead to greater synergy because of their complementary antibacterial actions. The analysis of correlations between the similarity and the antimicrobial properties allows us to effectively separate synergistic from nonsynergistic AMP pairs. Our theoretical approach can be used for the rational design of more effective AMP combinations for specific bacterial targets, for clarifying the mechanisms of bacterial elimination, and for a better understanding of cooperativity phenomena in biological systems.