The RapidCell (RC) model was originally developed to simulate flagellar bacterial chemotaxis in environments with spatiotemporally varying chemoattractant gradients. RC is best suited for motility simulations in unbounded nonfluid environments; this limits its use in biomedical applications hinging on bacteria-fluid dynamics in microchannels. In this study, we eliminated this constraint by coupling the RC model with the colloidal lattice Boltzmann (LB) model. RC-LB coupling was accomplished by tracking positions of chemoreceptors on particle surfaces that vary with particles' angular and translational velocities, and by including forces and torques due to particles' tumbling and running motions in particle force-and torque-balance equations. The coupled model successfully simulated trajectories of particles in initially stagnant fluids in bounded domains, involving a chemoattractant contained in a confined zone with a narrow inlet or concentric multiringed inline obstacles, mimicking tumor vasculature geometry. Chemotactically successful particles exhibited higher attractant concentrations near the receptor clusters, transient increases in the motor bias, and transient fluctuations in methylated proteins at the cell scale, while exhibiting more frequent higher particle translation velocities and smaller angular velocities than chemotactically unsuccessful particles at the particle scale. In these simulations, the chemotactic particles reached the chemoattractant with the success rates of 20-72 %, whereas nonchemotactic particles would be unsuccessful. The coupled RC-LB model is the first step toward development of a multiscale simulation tool that bridges cell-scale signal and adaptation dynamics with particle-scale fluid-particle dynamics to simulate chemotaxis-driven bacterial motility in microchannel networks, typically observed in tumor vasculatures, in the context of targeted drug delivery.