A sampling problem resulting in slow convergence of noise levels predicted by a Monte Carlo type low‐frequency underwater acoustics noise model, SIAM II, led to the development of a simple analytical ambient noise model, called CNOISE. The inputs to the model are azimuthal sector geometry about a receiver, ship counts in range‐sector bins, transmission loss curves for each sector, and ship source levels. The outputs available from the model include horizontally directional and omnidirectional ambient noise levels. The major advantages of this model include: Significantly increased speed relative to Monte Carlo noise models, and analytical solution that lends itself to straightforward analysis and implementation, and the avoidance of the sampling problems which appear to plague Monte Carlo models. Comparisons of noise levels predicted by other models (BEAMPL, SIAM II, and FANM) have shown very good agreement.