<p>This paper investigates the event-triggered tracking control for switched nonlinear systems (SNSs) with state triggering, unknown nonlinearities and parametric uncertainties. The main challenge of such a work lies in two aspects: (1) Unlike input triggering, the state triggering makes virtual control functions non-differentiable and thus undermines the backstepping control design for SNSs. (2) The concurrent switching of unknown parameters and mismatched nonlinearities fundamentally restricts the applicability of the multiple Lyapunov function methodology in SNSs. We solve the aforementioned problems by co-designing a new state event-triggered control strategy, which involves constructing first-order filters and integrating neural network and parameter projection methods. Exactly, two classes of state event-triggered mechanisms for system states and filter states are provided. With state triggering, the possible asynchronous switching phenomenon induced by input triggering can be avoided. The switching and mismatched nonlinearities are addressed by the neural network approximation and the uniform estimation of weight upper bounds. By developing mode-dependent control laws with parameter projector-based adaptive laws and event-triggered state signals, the practical tracking performance of SNSs is achieved and the switching condition is relaxed to the dwell time with an arbitrary positive value. To validate the obtained results, the tracking control of a networked single-link robotic manipulator is utilized.</p>