We proposed an extension of the ant colony optimization metaphor for continuous domain. This new approach was named Differential Ant-Stigmergy Algorithm and was studied on a set of new benchmark functions (defined in 2005) of real-parameter optimization problems. <br/> The algorithm was compared with a number of evolutionary optimization algorithms including Covariance Matrix Adaptation Evolutionary Strategy, Differential Evolution, Real-Coded Memetic Algorithm, and Continuous Estimation of Distribution Algorithm. <br /> The result obtained indicate a promising performance of the new approach. One can noticed that our approach perform better then the rest of the approaches on three out of four test functions. Since selected test functions reflected different kind of pseudo-real optimization problems, one can conclude that the Differential Ant-Stigmergy Algorithm is applicable to many real-parameter optimization problems. <br /> Regarding the future, one important issue consists of pure continuous ant-stigmergy algorithm. Here, so-called parameter differences will be in continuous form instead of fine-grained discrete form.