Conservation of soil moisture through tillage practices is an important management objective for crop production in semi-arid areas. A study to evaluate the effects of tillage practices on maize and bean yields was conducted in Mwala Sub County, Eastern Kenya, in the long (LR) and short rains (SR) of 2012/13. The tillage treatments were: Disc Ploughing (DP), Disc Ploughing and Harrowing (DPH), Ox-ploughing (OX), Subsoiling – Ripping (SSR), Hand hoeing with Tied Ridges (HTR) and Hand hoeing (H) only. There were three cropping systems of Sole Maize (SM), Sole Bean (SB) and Maize - Bean intercrop (M + B), which were investigated in a Split-Plot Design field experiment with four replications. Data on maize and bean yield and yield components were monitored throughout the four cropping seasons. Maize plant height, leaf area and leaf area index, maize and beans grain and biomass yields were significantly affected by tillage (P < 0.05). No significant effect of cropping systems on the maize height was observed. Higher maize grain yields (P < 0.05) were obtained in the sole maize plots in LR 2012 (5.01 Mg ha -1 ), SR 2012 (4.19 Mg ha -1 ) and in the SR 2013 season (2.82 Mg ha -1 ). A three - season bean grain yield average by tillage shows that DPH > SSR > DP > OX > HTR > H, with values ranging from 0.75 Mg ha -1 to 1.46 Mg ha -1 (P < 0.05). Intercropping reduced the seasonal means of bean grain yields (P < 0.05) with a 54 % decrease by intercropping (0.73 Mg ha -1 ) compared to the sole bean (1.6 Mg ha -1 ). Thus, the DP and DPH improved crop yield and yield components and can be recommended as tillage practices in the semi-arid region.