We present the first sample of 882 optically selected galaxy clusters in the Deep Lens Survey (DLS), selected with the Bayesian Cluster Finder. We create mock DLS data to assess completeness and purity rates, and find that both are at least $70%$ within 0.1$le z le$ 1.2 for clusters with $M_{200}ge 1.2times 10^{14}M_{odot}$. We verified the integrity of the sample by performing several comparisons with other optical, weak lensing, X-ray and spectroscopic surveys which overlap the DLS footprint: the estimated redshifts are consistent with the spectroscopic redshifts of known clusters (for $z>0.25$ where saturation in the DLS is not an issue); our richness estimates in combination with a previously calibrated richness-mass relation yields individual cluster mass estimates consistent with available SHeLS dynamical mass estimates; synthetic mass maps made from the optical mass estimates are correlated ($>3sigma$ significance) with the weak lensing mass maps; and the mass function thus derived is consistent with theoretical predictions for the CDM scenario. With the verified sample we investigated correlations between the brightest cluster galaxies (BCG) properties and the host cluster properties within a broader range in redshift (0.25 $le z le$ 0.8) and mass ($ge2.4times 10^{14}M_{odot}$) than in previous work. We find that the slope of the BCG magnitude-redshift relation throughout this redshift range is consistent with that found at lower redshifts. This result supports an extrapolation to higher redshift of passive evolution of the BCG within the hierarchical scenario.