This thesis addresses the problems in electronic structure methods for modeling chemical reactions, particularly those that involve nonadiabatic transitions between excited and ground electronic states. Accurately modeling such reactions (eg. photochemical reactions) requires methods that can treat the ground and excited states in a balanced fashion. As such, one would turn to multireference methods. However, these methods are expensive, especially when we have to model large systems and calculate energies and gradients at every step in a dynamics run. In contrast, DFT/TDDFT are very attractive in this regard because they can account for dynamical correlations while remaining cost-effective. Despite this, DFT/TDDFT fails in the regions of crossings between the S0/S1 states. In this work, we develop the recently proposed TDDFT-1D method, which is a semi-empirical extension of TDDFT that includes one double in a configuration interaction Hamiltonian. In our benchmarking, we find that this method can produce accurate vertical excitation energies as well as smooth potential energy surfaces even in the regions of S0/S1 crossings, outperforming standard TDDFT. We then present the derivation and implementation of analytical gradients and derivative couplings for TDDFT-1D, which are required for nonadiabatic ab initio molecular dynamics simulations. In the last part of the work presented here, we propose a novel way to restore angular momentum conservation in surface hopping algorithms. The standard surface hopping method should conserve the total linear and angular momentum when propagating dynamics. However, when the momentum is rescaled in the direction of the derivative coupling, the angular momentum is not conserved because the derivative coupling vectors calculated from standard electronic structure methods are not rotationally invariant. We work out a way to remove the rotational components of the derivative couplings in a size-consistent manner such that this method can be applicable to surface hopping algorithms moving forward.