The invention refers to a method for synchronizing clocks in a communication network, wherein a first clock of a first network element (MA) which is a master element is used for synchronizing second clocks of one or more second network elements which are slave elements. According to the method of the invention, a first sequence of first messages transmitted from the first network element to the second network element and/or a second sequence of second messages transmitted from the second network element to the first network element is recorded. First messages and/or second messages out of those sequences are identified by using an appropriate threshold function with respect to the transmission delays of those messages. Those identified messages have the same constant minimum delay, and based on those messages clock synchronization between the first clock and the second clock is performed. The invention has the advantage that clock synchronization is possible even if an intermediate switch causing an unknown delay is located in the transmission path between the first and the second network element. This is because most of the messages are transmitted within a minimum constant delay via such an intermediate switch and, by identifying those messages, a line delay between the first network element and second network element can be estimated and used for synchronizing the second clock with the first clock. The synchronization method of the invention is preferably used for synchronizing clocks in a DECT network. Furthermore, in a preferred embodiment, the first and second messages are messages according to the standard IEEE 1588.