In the current protocol stack for Internet ofThings in general and wireless sensor network in particular,many devices rely on the ContikiMAC protocol at their MAClayer. This protocol is widely used and enabled by default forseveral industrial environments and time sensitive monitoringand control applications. However, few work exists regardingthe performance of this protocol because it lacks of anunderlying theoretical model for analysing its performance.In this paper, we propose a novel approach relying on processmining technique that aims to obtain a Markov chain modelfor networks running the ContikiMAC protocol. In particular,we present a comprehensive specification of the protocol anda Markov chain model obtained through the analysis andinstrumentation of its reference implementation. We used theobtained Markov chain to analyze and estimate the end to enddelay distribution for a multi-hops transmission with staticrouting. The approach can also be extended to a wide rangeof protocols.