Massive penetration of renewable energy brings great challenges to the secure operation and reliable power supply of important users in urban power grids, and it is easy for the random fluctuation of renewable energy power to result in cascading failure risks. A cascading failure risk assessment method for an urban power grid considering the random fluctuation of renewable energy outputs and T-connected line switching in a 110 kV power grid is proposed. The influence of the probability distributions of system states is taken into account in the calculation of the failure probability and consequences, and the probabilistic power flow based on cumulants is used to reflect the relationship of the probability distribution characteristics between the system states and renewable energy outputs. Additionally, a mixed-inter nonlinear optimization model to calculate the minimum load shedding considering the T-connection line switching of 110 kV power grid is established. Minimum load shedding is used to characterize the consequence of cascading failures. By deciding the switching state of each group of T-connection lines under faults, the load shedding in cascading failures can be reduced, and the risk of such failures is effectively decreased. The secure operating range of bus voltage of important user loads is described by chance constraints to ensure that the probability of important user loads without failure meets the given confidence level, so as to ensure the secure and reliable power supply for those users. Finally, the correctness and effectiveness of the proposed cascading failure risk assessment method are demonstrated in an actual urban power grid.