One problem of intensive fish culture systems is the progressive build-up of toxic wastes such as ammonia. The possibility of improving aquaculture water quality using two kinds of zeolite is discussed. Zeolites are alumino-silicates whose framework allows them to exchange cations. Ion exchange has been demonstrated to be competitive with other methods of ammonia removal due to the high selectivity for ammonia exhibited by zeolite materials. In this study an unknown Zambian zeolite (identified as laumontite by X-ray diffraction techniques) and Pratley clinoptilolite (a South African zeolite) were tested under laboratory conditions and in a fresh water recirculating system. Ammonia cation exchange capacities (CEC) and suitable application rates for efficient water treatment were determined using the batch and column ion exchange procedures. Estimated ammonia uptake, the most important criterion used to assess performance of zeolite filters was strongly influenced by zeolite type, particle size, pre-treatment, regeneration and ion exchange method used. Statistical analysis showed significant differences in average ammonia CEC values between clinoptilolite (14.94 mg g⁻¹) and laumontite (2.77 mg g⁻¹), with the former displaying a higher Na⁺ ® NH₄⁺ exchange rate especially in the early reaction stages. This difference accords with the higher purity of clinoptilolite, 47% as opposed to 4.7% for laumontite, which makes it a better zeolite for ammonium removal. CEC increased linearly as particle size of the clinoptilolite was reduced resulting in a linear regression model (y = 18.29 – 3.704 x; r² = 74%). Pre-treatment of clinoptilolite using 1N NaCl significantly improved the ammonia CEC of clinoptilolite. Overall performance of both the batch and column methods achieved after regeneration (18.3 mg g⁻¹) was 25% higher than the estimated CEC values (13.0 mg g⁻¹) for the unregenerated samples of clinoptilolite. Comparison of CEC estimates using Pratley clinoptilolite, showed that average batch CEC estimates were significantly lower than the column method estimates. The average ammonia CEC values estimated in a fresh water recirculating system (5.80 mg g⁻¹ and 4.12 mg g⁻¹ for the 0.7-1.0 and 1.0-1.4 mm particle sizes, respectively) were significantly lower than the column and batch estimates for the same particle sizes (P < 0.05). Some nitrite (NO₂) and nitrate (NO3) build up was experienced probably due to the growth of autotrophs in the filters. Mass balance of nitrogen (N) for the three treatments of the fish trial (0.7-1.0 mm, 1.0-1.4 mm and the control treatment that had no zeolite in the filter) indicated that less that 10% of the N was retained for growth. It was found that 60% of the NH₄-N present associated with the soluble N was available for absorption by the zeolite filter or biological nitrification and that a total of approximately 22% of NH₄-N available was absorbed by clinoptilolite. The results indicate that the rate of nitrification can be deductively estimated by allowing a zeolite filter to become a biological filter. It is concluded that water treatment by ion exchange using natural zeolites, provides a reliable and efficient method for ammonia removal and appears to be a viable supplementary water treatment method for fresh water systems.