In this work, a new bio-inspired classification algorithm is proposed to solve offline handwritten signature verification. Precisely, we investigate the usefulness of the Artificial Immune Recognition System (AIRS) to allow an automatic detection of forged signatures. For feature generation, various topological and textural features that are pixel density, Longest Run Features, Local Binary Patterns, and the Gradient local binary patterns are investigated. Experiments are conducted on the CEDAR dataset, which contains signatures of 55 writers. The resulting error rates confirm the effectiveness of the proposed verification system since it overcomes several state of the art systems.