The problem of path selection when sending information from one node to another over multiple hops, solved by routing algorithms, is a substantial one in computer networks. Especially in unstructured Peer-to-Peer networks, the topic is of major importance, since no global view on the network or global address mapping exists. This thesis provides a much needed benchmarking framework that allows the fair and systematic benchmarking and comparison of routing algorithms in unstructured Peer-to- Peer networks. The resulting application, based on the Peer Model (a coordination based programming model), supports easy exchangeability of routing algorithms and extensive configurability. Additional contributions are the adaption of existing swarm intelligent algorithms from a different domain to the domain of routing. BeeNet is based on the foraging behavior of honey bees, whereas SlimeMoldNet makes use of the Dictyostelium discoideum slime molds life-cycle. Both algorithms are competitively benchmarked, evaluated and compared to five well known routing algorithms: AntNet, BeeHive, Physarum polycephalum routing algorithm, Gnutella Flooding and k-Random Walker. Overall, SlimeMoldNet outperforms the other algorithms in regards to the average data packet delay. This especially holds for bigger P2P network sizes and data packet traffic levels. BeeNet shows similar good results. In terms of scalability, BeeNet outperforms all other algorithms, beside k-Random Walker at some occasions, without having the same major drawbacks. SlimeMoldNets scalability is above average and improves drastically proportional to the network size and data packet traffic level.