The deterioration of track geometry depends on several factors of which the speed of the train is one. Imposing a speed restriction can slow down the track deterioration and allows a longer survival time before a serious condition is achieved. Preventive maintenance delays can be authorized during the survival time. However, speed restrictions also reduce the system throughput. On the other hand, a longer interval between preventive maintenance activities has a lower maintenance action cost and it also enables grouping the maintenance activities to save set-up costs as well as sys- tem downtime. If the repair delay is too long, it may cause unacceptable conditions on the track and lead to higher main- tenance costs and accidents. Therefore, it is interesting to assess the effect of a speed restriction on the delayed maintenance strategies for a railway track section. We want to solve a maintenance optimization problem to find the optimal tuning of the maintenance delay time and imposition of a speed restriction. To this aim, a delayed maintenance model is developed, in which track deterioration depends on the train speed and the number of passing trains. The model is used to determine an optimal speed restriction strategy and a preventive repair delay for the optimization of the system benefit and unavailability. Coloured Petri Nets are adopted to model the maintenance and operation of the railway track section. The Coloured Petri Net model describes the gradual track deterioration as a stochastic process. Different speed restriction policies and maintenance delay strategies are modelled and activated by the observed compo- nent states. Monte Carlo simulations are carried out to estimate the maintenance cost, the system benefit and the sys- tem downtime under different policies. Numerical results show the maintenance decision variable trade-off.