The performance of a reacting system for degradation of trichloroethylene (TCE) in a pilot-size annular photocatalytic reactor having a tubular lamp located at its centerline was studied. The reactor operation was carried out with commercial catalytic particles of titanium dioxide (Aldrich) in a water suspension. The description of the reactor performance was made by employing a kinetic model developed in a laboratory reactor of different size and configuration, irradiated with similar lamps of lower output power. The annular reactor was operated in the continuous mode but inside the loop of a recirculation system. The performance of the annular reactor was modeled by assuming three different behaviors: (i) a well-stirred tank reactor, (ii) a pseudo-steady-state laminar flow reactor in a batch recycle, and (iii) a transient-state laminar flow reactor in a recycle. Case iii produced the best representation of the experimental data. Along with the experimental validation of the developed theoretical models, three practical features were unveiled that had to be taken into account in any Subsequent design: (a) an important change in the pH, from 6.5 to 3.5, along the reaction time; (b) a significant titanium dioxide deposition on the reactor walls; and (c) an appreciable catalyst agglomeration produced by the recirculation system and magnified by the change in pH. Only after the first 120 min of reaction time, the reactor conditions became almost stabilized; thus, the observed transformations had to be incorporated into the modeling.