Chickpea is an important legume crop in the arid and semi-arid lands (ASALs). It is commonly grown during the post-rainy season under receding soil moisture conditions. This exposes the crop to drought especially terminal drought which occurs towards the end of the cropping period, causing high yield loses. Developing drought tolerant and high yielding chickpea genotypes is an important goal for plant breeders in the ASALs for increased productivity. The objectives of this study were to a) identify production constraints and farmer preferred traits b) determine the inheritance of root traits and yield component c) introgress drought tolerance root traits through marker assisted backcrossing, and d) evaluate genotypes and identify quantitative trait loci (QTLs) for yield components in chickpea. The study was conducted in Kenya during a three year period (2012 to 2014). Participatory rural appraisal (PRA) through focus group discussions established that the major chickpea production constraints were drought, pest infestation, late maturing varieties, diseases and lack of market. Farmers preferred chickpea that were high yielding, drought tolerant, early maturing and tolerant to pests and diseases. Farmers especially in the dry highlands needed varieties that could fit in the short duration period as they planted chickpea as a relay crop. Desi (brown seeded) types of varieties were preferred over Kabuli (white seeded) by farmers. Genetic analysis through generation mean analysis revealed that total root length, root length density, root dry weight, shoot dry weight and 100-seed weight were governed by additive genes. However, root dry weight was also controlled by non - additive genes; dominance, additive x additive and dominance x dominance interactions. Similarly, shoot dry weight was governed by additive x dominance while 100-seed weight was also controlled by dominance and additive x additive genes. Five genes controlled 100-seed weight. Introgression of drought tolerance root traits using marker assisted backcrossing (MABC) from donor parent ICC 4958 into two varieties (Chania Desi II and LDT 068) was achieved. However, low polymorphism from both simple sequence repeats (SSR) and single nucleotide polymorphisms markers was obtained among the parents. Two families (EUC-03-BC2F2-P22-1-2-1 and EUC- 03- BC2F2-P22-1-2-3) were significantly different for seed weight per plant (g) and one family (EUC-03- BC2F2-P22-1-2-1) for 100-seed weight from the recurrent parent (Chania Desi II). BC2F3 families were significantly different for root dry weight (RDW), shoot dry weight (SDW), total plant dry weight (PDW) and root to shoot dry weight (R/S) for Chania Desi II x ICC 4958 and R/S for LDT 068 x ICC 4958. However, 20 families had better root traits: root length density (RLD), rooting depth (RDp), RDW, and total root length (TRL), than their recurrent parents in both crosses. SDW was positively correlated with TRL, RLD and RDW which could be used as indirect selection criteria for root traits. Evaluation of 188 F3:5-6 28 genotypes from ICCV 05107 x ICCV 94954 under irrigated and rainfed conditions indicated that four lines namely; ICCX-060045-F3-P174-BP, ICCX-060045-F3-P76-BP, ICCX-060045-F3-P4-BP and ICCX-060045-F3-P146-BP had yields ranging from 1102 kg/ha – 1158 kg/ha, than better parent, ICCV 05017 (689 kg/ha), across the environments. This was 60 % higher than the best parent. Similarly, under irrigated conditions four lines had over 100 % yield increment than better parent. These were; ICCX-060045-F3-P174-BP, ICCX-060045-F3-P146-BP, ICCX- 060045-F3-P23-BP and ICCX-060045-F3-P62-BP. Three lines, ICCX-060045-F3-P188-BP, ICCX-060045-F3-P111-BP and ICCX-060045-F3-P4-BP had over 50 % yield increase compared to the better parent, ICCV 05017, under rainfed conditions. Positive significant correlations were obtained between yield and biomass, harvest index, seed weight and days to maturity. Quantitative trait loci (QTLs) using IciMapping for yield traits generated a linkage map spanning a total length of iv 335.04 cM with 45 simple sequence repeat (SSR) markers and identification of eight QTLs. Three QTLs for above ground biomass were mapped, one on LG 3 and two on LG 4 (8.67-32.4 % phenotypic variation expressed, PVE). Two QTLs for yield were mapped on LGs 4 & 6 (8.24-11.08 % PVE). One QTL each was mapped for 100-seed weight on LG 1 (12.19 % PVE), HI on LG8 (9.9 % PVE) and days to maturity on LG 4 (13.31 %). Breeding for multiple traits such as high yield, drought tolerance and pest tolerant chickpea, as per farmers‘preferences will enhance chickpea adoption. Given that most of these traits are controlled by many genes selection should be done in the later stages of a breeding programme, from F4 generations onwards, when plants are approaching homozygozity. Families developed through MABC that had higher mean root traits and yield compared to recurrent parents and lines selected under irrigated and rainfed conditions should be extensively tested for possiblerelease as commercial varieties. This will enhance chickpea production especially in ASALs. Further, validation of identified QTLs and its application through introgression will be useful in accelerating conventional methods through marker assisted breeding.