With growing concerns about travel demand management practices in overcrowded metro systems, it is considered that time-dependent pricing strategies are effective for dealing with the crowding occurring during peak commuting hours. In this study, a bi-level optimisation framework is used to consider the peak avoidance behaviour of commuters in the development of time-dependent pricing strategies. The behavioural sensitivity of commuters to pricing factors is investigated in terms of departure time and mode shift decisions based on a stated preference survey conducted in Beijing, China. The proposed bi-level programming model comprises a multi-objective optimisation model at the upper level and a nested logit-based stochastic user equilibrium model at the lower level. Based on an empirical case study of the Batong line in Beijing metro, nine optimal time-dependent pricing strategies are tailored by representative decision preferences, yielding up to 13.97% decrease in the peak ridership during rush hours.