By inducing double-strand breaks (DSB), it is possible to initiate DNA recombination. For a long time, it was not possible to use DSB induction for efficient genome engineering due to the lack of a way to target DSBs to specific sites. With the development of modified meganucleases and synthetic DNA binding domains, this limitation was overcome. Domains derived from zinc finger transcription factors or transcription activator-like effectors can be designed to recognise almost any DNA sequence. By fusing these domains to the endonuclease domains of a class II restriction enzyme, an active endonuclease dimer can be formed that introduces a site-specific DSB. Recent studies demonstrate that gene knock-outs via nonhomologous end joining or gene modification via homologous recombination are becoming routine in many plant species. By setting a single genomic DSB, the complete knock-out of a gene, the sequence-specific integration of foreign DNA, or the subtle modification of individual amino acids in a specific protein domain can be achieved. The induction of two or more DSBs allows for complex genomic rearrangements such as deletions, inversions, or the exchange of chromosome arms. The potential of controlled genome engineering in plants is tremendous. The recently discovered RNA-based CRISPR/Cas9 system as new tool to induce multiple DSBs or sophisticated technical applications, such as the in planta gene targeting system, are further steps in this development. At the moment, the focus still lies on the engineering of single genes; in the future, the engineering of whole genomes will become an option. This article is protected by copyright. All rights reserved.