Travel time estimation is a challenging task due to the lots of influential factors. The most pre-existing travel time literature assumes travel time follows a specified distribution. Their goal is to find the best fit distribution of travel time based on the observed travel times. Obviously, the role of some influential factors such as intersection signal times, demand variations, supply variations (changing network topology) etc., are neglected.  In this paper, the uncertainty of time and its variations which are consequences of numerous stochastic elements such as uncertain travel demand, uncertain driver behaviors, and random arrival time is modelled by calculating  the probability density function of travel time on each arterial link. Also, the concept of queuing theory is used to take the stochastic features and queue formation behind intersection signals into account;  Additionally, unlike previous model-based methods, the queue of vehicles is modeled as a horizontal queue, and its effects on the travel time are accurately calculated. Also this method is particularly useful  in scenarios  where, due to changes in the network, the previous data may not be highly reliable or in special situations which there is only an estimate of supply and demand such as, the closure of a street or changes in traffic signal timing or large sporting events or emergency situations etc. Eventually, the proposed model is validated and examined by comparison with micro-simulation experiments.