We present an optimised galaxy cluster finder , 3D-Matched-Filter ( 3D-MF ) , that utilises galaxy cluster radial profiles , luminosity functions and redshift information to detect galaxy clusters in optical surveys . This method is an improvement over other matched-filter methods , most notably through implementing redshift slicing of the data to significantly reduce line-of-sight projections and related false positives . We apply our method to the Canada-France-Hawaii Telescope Legacy Survey Deep fields , finding \sim 170 galaxy clusters per square degree in the 0.2 \leq z \leq 1.0 redshift range . Future surveys such as LSST and JDEM can exploit 3D-MF ’ s automated methodology to produce complete and reliable galaxy cluster catalogues . We determine the reliability and accuracy of the statistical approach of our method through a thorough analysis of mock data from the Millennium Simulation . We detect clusters with 100 % completeness for M _ { 200 } \geq 3.0 \times 10 ^ { 14 } M _ { \odot } , 88 % completeness for M _ { 200 } \geq 1.0 \times 10 ^ { 14 } M _ { \odot } , and 72 % completeness well into the 10 ^ { 13 } M _ { \odot } cluster mass range . We show a 36 % multiple detection rate for cluster masses \geq 1.5 \times 10 ^ { 13 } M _ { \odot } and a 16 % false detection rate for galaxy clusters \gtrsim 5 \times 10 ^ { 13 } M _ { \odot } , reporting that for clusters with masses \lesssim 5 \times 10 ^ { 13 } M _ { \odot } false detections may increase up to \sim 24 \% . Utilising these selection functions we conclude that our galaxy cluster catalogue is the most complete CFHTLS Deep cluster catalogue to date .